CA1329953C - Polygonal information encoding article, process and system - Google Patents

Abstract

ABSTRACT
The article of the invention is an optically readable label for storing encoded information, said label comprising a data array of a multiplicity of information-encoded polygons arranged in a predetermined geometric pattern, and said poly-gons having at least two different optical properties.
A process for encoding information in an optically-readable data array comprised of information-encoded polygons by assigning optical properties to individual polygons in a predetermined pattern, ordering the polygons in a predetermined sequence, and printing the polygons with at least two optical properties.
A process for retrieving information by optically scan-ning a data array of information-encoded polygons, preferably hexagons, creating an optical replica of the digital bit stream representative of the optical properties of the information-encoded polygons, decoding that optical replica and retrieving the decoded bit stream.
A system for performing the foregoing encoding and decoding processes.

Description

^` 1 3 ~ 3 Microfiche Appendix A Microfiche Appendix is included in the present ap-plication comprising one microiche and a total of one test target frame and 78 ~rames of computer program listings.

Thi invention relates to an improved optically read-abla label lnd a reading 8y tem therefor, and, in particular, to an improved optically readablo label, attached to or printed on a ~ub~trat~, for storing inormation within a two-dlmensional data array, comprising a multiplicity of polygon~ arranged in a predetermined geometric pattern, and said polygon~ having at least two diferant optical propertie~.

Merchandise, various component part~, letters, pack-ages, containors and a whole gamut o relat~d item~ being Rhipped or transported, fr~quently are rsquired to be identl~iad with information as to origin, flight numbsr, de~tination, name, price, part num~r and numerous other kind3 of information. In other application3, reading encoded information printed on labeIR a-ix~d to 3uch items permlts a~tomation o sale~ ~igures and in-ventory or tho operation o~ olectronic cash re~lster3. Other application~ or ~uch 0ncoded 1abel3 i~clude tho automated rout-ing and sortin~ of msil, p~rcels, baggag~, and the like, and the placin~ o~ label~ b~arin~ manufacturing in~truction3 on raw mate-rial~ or Gomponent part~ in a manufacturing proce3~. Labels for the~e typo3 o~ artlcle~ are conventionally marked with bar codes, one o~ which $3 tho Univsr~al ~roduct Code. Numerous other bar code system~ are al o known in tho art.
Co~mercially-available bar cod~ sy~tem~ typ1cally lack ~ufficlent data den~ity to accommodato th~ pre3ent and increasing need to encode mora and more information on label~ of increas-lngly ~maller size. Attempts to raduce ~he overall siza and spacing o~ bars in various bar code ~ystems to increase data .` ' ' ' ' ` I . I

11 32~53 density have not qolved the problem; optical 3canners having suf-ficient r~solution to detect bar code3 compriqlng contrasting bars spaced five mll~ or less apart are generally not economic-ally feasible to manufacture because of the close tolerances in-herent in the label printing procasq ancl the sophisticated op-tical apparatus required to resolve bit~encoded bar3 of these dimensions. Alternatively, to accommodate incraased amounts of data, very large bar coda lab~lq must bi~ fabricat~d, with the result that such labels ar~ not compact enou~h to it on qmall articles. Another important factor i~ the cost o tho labal mcdium, such as paper. A small label ha a sma}ler paper cost than a large label; thi8 co~t is an important factor in large volume operation3.
Alternativeq to bar code~ include: circular formats employing radially dlspo~ed w~dge-shaped cod~d elomenta, such as in U.S. Patent 3,553,438, or concentrlc black and whlte bit-encoded rings, ~uch as in U.S. Patent~, No3. 3,971,917 and 3,916,160; grid~ of rows and column~ of data-encoded squares or rectangle , such a~ in U.S. Patant No. 4,2~6,146; microscopic ~ 20 ~pots dl~po~d in c0118 forming a regularly ~paced gri~, as in ; U.S. Patent No. 4,634,850; and den~ely packed multicolored data ', ~ields o~ dots or slemonts, ~uch as describod in U.S. Patent No.
i 4, 4~a, 679 . Some of the coding ~y~tems de~cribod in the foregoing ;' exampl~ and o~har coding ~yatem~ known in the ar~ primarily suf-I
er ~rom d~iciencie~ in da~a denslty, such a~ in t~e cas0 of encoded circular patterns and grids of rectangular or square boxe3. Alt~rnativoly, in the caq~ of the grids comprised of micro-scopic spots or multicolored elements referr~d t~ above, quch sy~tem~ require speci~.' orientation and tran~port maan3, thus limiting tholr utility to highly controll~d readlng ~nvironment-q.
~r~ D~o to th~ ~izo and sp~ed of modorn conveyor ~ystem~, (utilizing conveyor belt w$dths of 3 to 4 oet, for example) and having b~lt speeds approaching 100 inche~ p~r second or more, carrying packages o~ varying haightg on wh~ch inormation encoded labels are af~ixed, and the need ~o utilize a sm~ll, inexpensive, i . .

~ 132~3 compact label of about one ~quare inch, great strains are placed on the optical and decoding systems required to locate and read the data encoded labal~ on thqse rapidly moving package~ and the lik~. There a~e di~icultieY in the optical scanner simply acquiring the label image. Furthermora, once acquired or id~nti-fi~d, the label image must be accurately d~coded before the next operation on the pacXage in the conveyor sy3tem takes place, oftan in a raction of a second. These problems hav~ led t~ the need or pro~iding a ~imple, rapid and low-co9t mean3 of ~ignal ing the presencQ of a data-encodad label within the ~i~ld of view of an optical scannar mounted in a manner to p~rmit scanning the entir~ conveyor helt. This eature desirably i coup}ad with a high den~ity data array, described in mor~ detail b~low.
Data arrayo containlng acqui~itlon target~ are known in the art; for example, concentric geomotric figures, including rings, squares, triangles, hoxagons and numerous variations thereof, ~uch as de~cribed in U.S. Patent~ Nos. 3,513,320 and 3,603,728. U.S. Patents Nos. 3,693,154 and 3,801,775 also de-~cribe the uso o~ symbols comprising concantric circle3 as iden-tification and position indicator~, which symbols aro affixed toarti~las to be optically ~cannod. ~ew~v~r, thcse ~y~tem~ employ two ~eparate ~ymbol~ to d~termine the ids~ti~ication of the data ~.
i~1d and its po~ltion, thereby lncroa~ing tha complexity of the~
logic circuitry roguirad to det~ct the ~ym~ols, a well as reduc-ing th~ data-carrying capacity of the associated data field.
Al30, when t~o ~ymbol~ are usad, damago to one cau~es problems in locating tho position of ths data ield and ~he a~tendant ability to recover information from tho data ~ld. In the lattsr ystem, ~ap~rate posltion and ori~ntation markinsq ar~ utiliz0d at opposite 30 ~nd~ of data track~ having data-encod~d llnoar marXings of only limit~d data carrying capability.
The foregoing systems are generally qcanned with an optical ~n~or capable of generating a video slgnal output corres-pondi~g to the change in intensity of light reflectad off the data array and po~ition and orientation symbol~. The video out , ' 132~9~3 ,, , put of such syqtems, after it is digitized, has a particular bit pattern which can be matched to a predetermined bit sequence.
These sy~tems, however, su~fer the drawback of requirlng two separate symbolq for first ascertaining the image and secondly S d~termining lt3 orientation. Also, ~he proce~ of having to match the digitized signal outp~t o~ the optical qensor with a prede-termfned bit ~equenca representing both the po~ition and orienta-tion symbol~ more likely to produce erroneous readings that tha proce~o and system of thi~ invantiorl, becau~e the prior art label acqui~ition sy3tems provlde an inflexible charact~rization of the acquisition target signal level.
U.S. Pat~nt No. 3,553,438 disclo~es a circular data array havlng a csn~rally-located acgui~ition target compri3ing a 3erlec o concentric circlos. rh- acgui~ition target provides a msans of acquiring the circular iabel by th~ optlcal sensor and detormining its geometric center and thereby the geometric center of tha circular data array. This ls done through logic circultry operating to recognize the pul~e pattorn repre~entative o~ the bull~-~ye configuration of the acquiqition target. ~owever, as for har codos, ~h- data array has only a limltod data capacity a~d th~ sy~tem roquires second circular ~canning proces3. Use of both a linear and circular qcan for a sy3tem of such limited ~
data capacity cr~ata~ unde~irablo complexity in the 3ystem for a~ -slight gain in data capacity over conventional bar code3.
To increa~e the data carrying capacity of data arrays, codes employing multiple high donsity colored dot~ have been d~volop-d, as doscrlbed in U.S. Paton~ N~. 4,48 ,679. Sy3tems o i th~ typo do~rib~d ln U.S. Patent No. 4,498,679, howover, require th~ u~ o~ hand-hold optical ~cannera, which are totally incapable of recording and decoding rapidly movinq data ~rrays on a pac~age being transported on a high~speed conveyor belt. Analoqously, high density coding Ryqtem~ employing ~icro~copic data-encoded spots, as de~cribed in U.S. Patent No. 4,634,850, require special transport mean~, thereby en~uring that the data array is moved in a speci~ic direction, rather than ~imply at a random orientation, ; -5-, 13299ri~
as might be found with a package being transported on a conveyor belt or the like. Thus, the coded label must be read track by track, utilizing a linear scanner coupled with label transport mean3 to properly decode the lnformation encoded on the label.
Also, in this patent, the position of ths card in relatlon to the sensor mu~t be very carefully controlled to be readable.
Multiple color~ have al~o been utilized in the art of producing bar code system~ 30 as to overcome the optical problems of scanning very minute bar~. A bar code uti}izing more than two optical propertie~ to encode data in a data array, by for in~tance, u~e o~ altornating blac~, gray and white bar3, iq described in U.S. Patent ~o. 4,443,694. Howover, systems of the type described, although an lmprovement over earlier bar code ~ystems, neverthe-les~ fail to achieve the compactness and data den~ity o:E the invention described herein.

In view of the foregoing drawbacks of prior optical coding sy~tems, it i~ a princlpal object of this lnvention to pro~ide new and improvsd compact, high-lnformation-density, optically-raadablo labels.
Another ob~oct of the invention i~ to provide new and improved optically readablo label~ which may be encoded with abo 100 highly error-protected alphanumeric characters p~r square lnch of labql aroa.
Still anothor objoct of thi~ lnvention 19 to provide new and improved compact high-information-den~lty, optically-readable labels which may~be read by an optical sensor when the label is affixed to a package or other like a;~icle boing tralls-ported on a high speed conveyor system, without regard to theorientation of the-package thereon or the variabllity of the heigh~s o ~aid package3 upon which the optically-readable label is a fixed.
A concomitant object o this invention 19 to provide an optically-readable label and decoding system combination, so t~at , ' ~ -- .
~ ~, ~ ~2~953 the label is capable of being reliably decoded eve~ if tilted, curled, warped, partially obliterated or partialLy torn.
An add$tional object of thi invention i3 to provide methods for deter~ining the location of a label pa ~ing under S an optical sensor at high speed and decoding said label with a high degree of data integrity.
I~ is a further ob~ect o~ this invention to provide improved methods of encod~ng compact, high-information-density, im~roved, optically-readable labels by dividing the information to be encoded lnto h~gher and lower priority mes~age3, to create a hierarchy of me3sages, which are ~eparately error protected to en~ure the integrity of the encoded informa~lon.
It i yet another object of this invention to provide improved m~thods and sy3tem~ of encoding and decoding compact, high density, improved, optically-readable labeL3 which include error correction capabilities, ~o as to ra~tore misread or miss-in~ lnformation and to do so with a preferenc- to the high prior ity encoded mes3age.
A furth~r ob~ect of thl9 i~ventlon is to produce in exp~nsiv~ optically-roadable labels by conventional printing means and decodlng 3ame, with relativqly ine~p~n3ivs logic circui~ry.
Furth~r objects and advantagas of the invention will becomc apparent from the d~scription o tha invention which follow~.

Tho pro~ent invention comprlses an optically-readable label for -~toring data encoded in bit orm, comprising a predeter-min~d two-dim~nsional data array o a multiplicity of information-encoded polygons arrangad contiguously, partially conti~uously ornonccntiguously in a pr~datermined two-dimen~ional pattern and ha~ln~ at laast two diferant op~ical propertios 2S well as m~thods and apparatus or encoding and d~coding such optically-readable label~.

, 1 3 ~
Optically readable label~ of the invention may comprise, predetermined two-dim,~,nsional geometric arrayq of polygons where the ~3,,0metri~ centers of such polygons lie at the Yertices of the intersecting axes as more fully discu3sed below of a predetermined two-dimensional array and wherfo, the polygons ha~e one of at least two diferent optical properties. The polygons o such optically readable labels may be regular or irregular polygons and the two-dimen3ional arrays of polygons on the optlcally readable label~
may have two or more egually- or unequally-angularly spaced a~eq in thf~3 plano of ths label.
Optically roadable labfels may be printed with con~lg-urations of polygons which are totally contiguouç~, partially contiguous or noncOntiguoUB. The latter two coniguratlons in-herentLy deine a multiplicity o interstitial spaces on the optically readaible labsl between ad~,ace~t polygons. Such in-ter3titl31 ~paca~, may havfo the i3ame or di~er~,nt optical prop-erties as the tWo or more optical propf~3rties of the polygons.
Two-dlmf~,,nslonal arrays o~ contiguoui3 polygoni3 having five or more isidss arf~3 usable as optically rfe,,adable label confi~uration3 o~
the invention. Al30, two-dimo,nsional arrays O~ either ragular or irregular, and e,ithffsr partlally contiguoua or noncontif~uous, polygons ha~ing thr,3fc~, or mora sid~efs, when prearranged on f.
predetf3rminsd axes of i~fuch array3, may be encodad and decoded in accordance with thfoff processe3 of the invention.
! 25 In addition to tha ~oreqoing varifftiff~s of geometric -i polygonal cell~, arrangomfsnts of such polygonal cell~, and geom~i etries o~ th0 optically raadable Iabel~ formed by ~uch arrange~
` menti~ of polygonal cf~Sflli~ hf3 optically readabflf~f labole~ of the ;:
,~ inventlon may f~ifptionally contain an acf~uisition target comprising a f~erlf~f,~ o~ concsntric ringf~, t4 aid ln the locat~ng o~ the optically readabla lab~l on tha articlf3,a upon which they are a fixed, particularly in dynamic lab,31 reading systems.
In a pr,~f,~f3,rred e~bodimen~ of the invantion, the data ~, array compri~es a generally isf~uare~shapad array of about one '' 35 square inch, having contiguously-arrangad hexagons orming rows , i~ -8-,~ . ;-"
,, .

~ 3~9~
and columns and a centrally-located acquisition target having a geometric center which defines the geometric center of the data array. The acquisition target may be any of a number of geomet-ric shape~ having optical properties capable of generating an easily recognlzabla video signal when scanned with an optical sensor across a linear scan line pa~sing through the geometric center of tha ac~ui~itio~ target. In a pre~erred e~bodiment, the acquisition target is a plurality of concentric rings o contrast-ing reflectlvities, which will yield a periodic vldeo signal when ~canned linearly. By using analog Æilter means as part of the m~thod oi' locating and decoding the data array, the signal generated by the optical sen~or is comparad directly with a pre-determined frequency, thereby allowing rapld and preci e matching o the fre~uencie~ and con~equent d~termination o the location of the data array affixed to a ~ub~trate. Tho analog slsctrical signal output from the optical ~ensor reprosenting the lnformation-ancoded label i9 then digltized and decodod. Utlllzing an analog bandpa~ filtaring step permits labsl acqui~itlon to occur without the need for decoding the information-ancoded label. By locating the cent3r o~ ~he acguisition targat a reference point on the data a~ray may ba determined. If the c~nter of th~ acquisition ta~got io locat~d at the c~ter o th~ lab~l, a ~imultan~ou~ deter~
mination of tho c~ntor o~ the acquiRition target and the data array m~y be accomplished. A central location o~ the acquisition 2S targ~t on the label i9 pr~erred, but not required, in the practice o~ the sub~ec~ invention.
The optically~x~adable data array o the present inven-tio~ i4 capabl~ of encodlng 100 or up to several hundred or more ~rror protected alphan~moric characters in an area of about one 3guar~ lnch when hoxagon~ ara ~ncodsd u~ilizin~ thrg9 r~flective properti~s, such a~ the color~ black, whit~ and gray. For a ~en30r wi~h a given optical resolution, the sy tem o th~ inven-tion permits a much denser information packing capabili~y than is po~3ible with bar code sy~tem~. For example, if a high r&solution optical ensor i9 used with the system of this invention, hundreds ~--` 1329~3 of alphanumeric characters may be encoded in a ~quare inch.
Alternatively, 100 character~ per ~quare inch may easily be de-tected with a relatively low resolution sensor with the system of thi~ invention.
Optically-readable labels o the present invention may be produced with varying data densities by utilizing as ~ew aq two or more contra~tinq optical propertie~. Greater data densi-ties and the incluqion o~ an acguisitlon target in the ~ystem o~
this invention require a scanning apparatu~ of increa~ing com-plexity and the addition o more elaborate decoding algorlthms to read the encoded mes~ge, when compared with a bar coda rsading ~y~tam.
In this invention, data encoding may bc accomplished by encodlng a plurality oi~ blt3 from a blnary blt ~tream lnto a clu~ter of contiguou~ hexagons, each hexagon having one o at lea~t two optical properties, aLthough the oncoding coùld alter-natively b~ done on a hexagon-by~hexagon ba31s. Th~ digital bit tream may be generated by a computer, based upon data entered manually or otherwl~Q con~arted into a binary bit ~tream, or may be p~ovided a3 a pre~ocorded digital bit tream. The data to be encoded is bit-m~ppod ln a pred~terminod 4eguenco and within pre-detormined geographlcal aroas of th~ data array ~o increase the number of tran3itions bat~ean hexagon~ having diferent optical prop~rtl~s .
In the preerred embod~m~nt of the pres~nt invention, the messages to be encoded are divided into high and low priority message3, w~lch ara ~eparately mapp~d ln different geographic ar~a3 of th~ data array. The high priority mes-qage may option-ally be d~plicated in th# low priority mossago ar~a to reduce the 30 pos~ibility o losing the high priority m~sage due to Ycanning error caused by smudge~, tear~, old~ and other types of damage to the data array. The hlgh priority message iq ancoded in a central ar~a o~ the data array, near the acguisition target con-tained in the preerred embodiment, ln ord~r to protect the mes-~age rom damage which i~ mora likely to occu~ to the peripheral . ' 13~9~
, areaq of the data array. Error correction capabilities are desir-ably incorporated in the data array, utilizing the large in~orma-tion-carrying capacity of the present invention, to ensure a very high degree o data integrity upon decoding the message.
S In practicing the invention, a pixel grid of sufficient density to print the label with hexagon3 of different optical propertie~ i3 utilized, although alternative printing processes may be u~ed without departing ~rom tho spirit of the inventlon.
The pixol grid is bit-mapped so that, when the label is printed, tha optical properties oi each hexagon are predetarmined, o that they may later be decod6d to récover the data de~ignated by the encoding of th~ tndividual hexagons. ~hi~ typo o printing pro-ce~g i9 well known in the art and ~tandard printers and bit map ping technlques may be used to prlnt hexagons having the optical propertias required by thi9 invention.
Th~ present inv3ntion provide~ a new and improved pro-cess for retrieving the data encodcd in the bit-mapped array of polygons, preferably hexagons, forming the data array. Encoded lab~l9 may b~ pas3~d through a prodetermined llluminated area and opt~cally scannod by meanY of an electronically op~rated optical sensor or a hand-h~ld scannar may be pa~3Qd ovor the label~. The optical sensor produce~ an output which i~ an analog electrical sig~al corr~ ponding to ~he inten~ity of the indlvidual re1ecti property o an area of a label, as recorded by th~ individual pixel3 o~ the optical ~ensor. ~y mean~ of an analog filter, the analo~ ~ig~al o~ th~ optical 3en~0r i~ flrst comp~rod to a prede-torminet fr-quoncy valu~ corr~ponding to that o~ a prodetarmined acguisitlon target i~ lt i8 pr~ent on th~ data array. Once a ' good match is found, tha labeL i9 acquired and the center of the i 30 acqui~ltion target i~ deter~ined, thereby alJo determlning a refer- -~nce point on th~ data array. The analog signal i~ simultaneously digitized on a continuous basi~ by mean3 of an analog-to-digital convexter and 3tored in an image buffe~. Tha stored digitized data representing ~he entir~ labsL is available Por further pro-cessing in the decoding process.

.

: ' :~ . . .
:, : : : : .
.

~3~g~
, .
By ~tored program logic circuits, the digital data istransformed into a map o~ the interface~ o5 hexagons having dif-ferent optical properties. In a preferred embodiment of the inventlon, thi3 i~ done by computing the standard deviation of the inten~ities of the reflective propertie~ recorded by the op tical sen~or at each pi~el and a predetermined group of pixels urrounding that first pixel. ~igh ~tandard deviations therefore corre~pOnd to tran~ition area~ at the interface~ of contrasting hoxagons.
F~rther data transformation3, involving filtering pro-grams to determine orientatlon, direction and spacing o~ the hexa-gon~, ars performed on the digital data. The q~neral 3tepY of thi~ proce~ ar~:
(l) Filterin~ tha non-linear transformed ver~ion of lS the digitized lmage.
(2) Detormining the orientation of the label, prefer-ably by ~ocating the threo axes of the image (as iLlu3trated in Fig. 2) and determining which axl~ i9 parallel to two 3ide~ of the label.
~3) Eind~ng th~ center o each hexagon and determining the gray l~vel at each center.

(4) Tran~forming the gray level~ to a bit 3tream.

(5) Optionally, applying error correction to that bit stream; and ~6) Op~ionally, converting the bit ~tream to a pre-datormined s~t o characters.
It i3 to bo not~d that, although the proces~ of this in~ention is do wribet a~ applied ~o hexagon~ havlng two or more sptlcal prop~rtie3, the proces3, i~ partlcular, the steps for ad~usting tho optical lmage for label warp, tear and the like, may be applied to other type~ of label3 and other polygonal cells.
Other obi~cts and f~rther scope o applicability of the pre~ent invention will b~com~ apparant from th~ De~ailed Descrip-tion of the Invention. It is to be understood, however, that the detailed description of preferred embodlments of the invention is .

:
, ~32~9~
,., ~
given by way of illustration only and is not to be construed as a limitation on the scope of variations and modifications falling within the spirit of the invention, as made apparent to those s~illed i~ the art.
S

FIG. 1 is a plan view of an acquisition target o concentric rings in accordance with the present invention.
EIG. 2 is a fragmonted plan vtew of an optically-readable label having contiguou ly-arranged hexagons ~or encoding data in ac-cordance with the present invention.

FIG. 3 is a pian ~iew of a complete optically-readable label hav~
lng contiguously-arranged hexagonq of three optical properties for encodlng binary data and including an ac~uisition target, in acco~danco with this invention.
FIC. 4 i~ a plan viaw of a thres cell by thr~e call cluster o contlguou~ h~xagons, which may s~rve as the basic encoding unit o~ th~ pr~erred embodiment of thi~ invention.
FIG. S is a clu~ter map howing a graphic repra~entation of a data-array comprising 33 row~ and 30 columns, orming a grid of ~ -11 row~ and 10 colum~ o three cell x three cell cluster coding units of hexagons.

FIG. 6 i3 a ~chematlc vlew of a camera ad~usting system in ac-corda~ce with the invention for adjustlng ~he position of the optical light 3en~0r in accordance wi~h the height of package being qen~ed.
FIG. 7 is a detailad outline of th~ decod$ng proces~ of thls inventi~n.
FIG. 8 i~ a flow chart ~how~ng the acgulsition tarqet location ,~, process.

.

,. 1329~3 FIG. 9 is a flow chart showing the encoding and decoding program structura and data flow.

FIG. 10 i~ a flow chart showing tha se~lence of ima~e proc~ssing ~tep_.

FIG. ll i~ a plan view of a cluster of c:ontiguou3 regular hexa-gons arranged with the geometric center~; of ad;ac~nt hexagons lying at the vertices of a regular hexac30nal array.

FIG. 12 is a plan view of a clust~r of contlguous irregular hexa-gon~ arranged with the geometric center3 of adjac~nt hexagons lying at th~ vertice3 of an irregular hoxagonal array.

FIG. 13 i~ a plan view of a clu~ter o partially contiguouq poly-gons sub0tantially in the form o he~agon~ arrangod with th~ geo-metric center~ of adjacent polygon~ lying at the vertices of a hexagonal array.

FIG. 14 is a plan vi~w of a cluqter of contiguou polygons sub-qtantially in th~ ~orm of hexagon arranged with th~ geometric centers o~ ad~a~ont polygon~ lylng at the v~rtices of a hexagonal array.

FlG. 15 is a plan view of an optically readabLe label having con-tiguou3 polygon~ sub~tantially in th~ fo~m of hexagons arranged with th~ gaometric centers of ad;acent polygonl lying at the v~rtices of a hexagonal array and including an acgui~ition tarqet in acGordanco with thi~ inv~ntion.
.

EIG. 16 i~ a plan YieW of a cluster of contlguous~gullateral ~guaras arranged with the geo~etric centers of adjacent squares lying at ~he ver~lces o~ a hexagonal array.

~ 132~3 FIG. 17 is a plan view of a cluster of noncontiguouq rectangleq défining interstitial spaces among said rectangle~ with the geo-metric centers of adjacent rectangle3 lying at the vertices of a hexagonal array.
S
FIG. la i3 a plan viaw of a cluster o noncontiguou3 pentagons de~ining interstiti~l spacas among said pentagons with the geo-metric center~ of adjacent pentagons lying at the vertice3 of a h~xagonal array.
FIG. 19 is a plan view of a cluster o~ contiguou~ rectangles ar-ranged in stagqered rows and cOlUmn9 with the geometric centers of ad;acent rectangle~ lylng at the ver~ices of a hexagonal array.
FIG. 20 is plan vlew of a cluster o~ partially contlguous octagons defining lnterstltlal space~ among said octagons with the geometric centers of ad~acent octagons lying at the vertices of a rectangular array.

~.

' ; . ` ' '' ' ' ~ ', ~ ' ' `

~ ~329~3 The ability to encode information by virtue of the con-trasting colors of contlguout hexagonq or "c2119" arranged in a honaycomb pattern in a predetermined se~luence and array permits the information stored on the label ~o be recovered by an electro-optical sensor. Polygonal cell~, other than hexagons, that are arranged w1th the geometric centers of ad~acent polygons lying at the verticeY of a hexagonal or other pr~detormined array, may likewlse be u~ed to encod~ information on an optically readable label. Such polygonal ccll , whsn arrayed with t~eir respective ~ :
centers at predetarmined locations on a two-dlmen~iona} geometric array and when encoded in a predetermined sequenco, through as-signlng differe~t optical properties to a plurality of such poly-gonal cells, may be "read" by an electro-optical Jen30r and sub-~eguently decoded in accordance with tho proco~s of the invention de~cribed bolow.
~ he polygonal cell3 o the tnvention are information encoding unlts formed by a closed brok~n lin~, whlch cells are arrayed ln a predetermined two-dimenYional pattern on an opti-cally readable label. Label con~lguration3 employing a wide variety o~ polygonal ~hape~, and arrays of varylng geom~tries, such as hoxagonal, rec~angular or square array3, are usable in th~ practic~ of the inventlon. "Ad~acent" polygonal cells may be totally contiguou~, partially contiguous or nonco~tlguous on the optically r~ad~bl~ label of tho inV~ntton.
nContlguou~ polygons" ar~ polygonq arrang~d with the geometric c~nter~ of ad~ac~nt polygon~ lying at the vertices of a , pr~determinod two~dimensional array and with the borders of such `I polygo~s touchin~ the borders of immediately ad~acent polygons, leaving no inter~titiaL spaces. "Partially contiyuou~ polygons"
aro polygon~ arranged with the geometric centers of adjacent polygons lying at the vertices of a . ~ . , -: :

~ .

predetermined two-dimensional array and which polygons are sepa-rated Jomewhere along their respective borders from other surround-ing polygons, thereby cau~ing a multiplicity of interstitial spaces to be interspersed among 3aid polygon~ on the optically readable label. "Noncontiguous poiygon3'l are individual polygons arranged with the geometric center3 of adjacent polygons lying at the vertice~ o a predeterminsd two-dimen~ional array, and having no contact between the borders of an individual polygon and poly-gons 3urrounding aid polygon. Additionally, the polygonai cells and tha predetermin~d two-dimensional grids or arrays upon which the cont~r~ of ad~acent polygons are located may bo irregular, having u~ogually-3pacsd axes, or regular, having equally-spaced axes" in configur~tion. Such two-dimen ional array axes may be ~:
independant o the axes of symmetry, if any, o the polygonal 15 cells.
A~ used in the label o~ thi~ inventlon, haxagons pre-sent certain lmportant advantag~ Xor encoding inormation on a label. Tho~e a~vantages are:
(1) Eor a givon opticai re~olutlon, hexagons can ~e more densely pack-d than other polygons. For example, at a given re~olution, th0 corn~r~ o square~ ara more diicult to resolve, 90 that otheswise unneces~ary optical resolution is required to "r~ad" ~quar~. Clrcles would be optimal for optical resolution, but the pac~ between adjacent ~ircl~s would be wa~ted and would complicate tho proce3sing and printing of t~e labeL imzge, be-cau~e o~ ~h~ neod to aaslgn an optlcal propssty to the spacs~.
~exagon~ p~rmit optl~um pac~ing oi lnSormation, compared with clrcloq or other polygons lncludin~, octagons, ~guare~, triangles and the liks. Sguare~ an~ triangLes are problem~ because of the ~harp corn~r~ they hava. Clrcles and octagonq are problems be-cause of the wa~ted ~pace ~etween adjacent circles or octagons.
(2) A grid of contiguou3 hexagons has three axes. By using a lab~l o~ a sguare or rec~angular shape the major axis of the hexagon can be located by its predetermined relatlon to a side of the label. This location of ~he major axiJ of a hexagon .. . . . .

.: : .

~32~3 .,.~ .
grid facilitate~ the reading of the data encoded in the hexagon by its relation to that major axis.
A~ u~ed herein, "label" inc}ute~ a discrete unit, with a ~uitable adhe~ive backing, to be attached to a package or prod-uct, and ~he exterior surface of a container or other object onwhich optically-readable information i5 imprinted in accordance with thi~ invention.
A~ u~ed herein, "optically-readable data array" or "data array" mean~ a pattern o contiguou~ he~agon3 or other polygonal calls having two or mor~ optical prop~rtie to ~ncode, in retriev-able orm, a body o~ data by virtue of the re~pective optical prop~rtieq of and spatial relation~hip of th~ hex~gon~ or other polygonal cell~ to each other. The hexagons or polygon~ imprinted to contain this recoverable information ara re~erred to herein as ;'inormation-encoded" hexagon~ o~ polygons; because of the mannsr in which the label encodes in~ormation.
The pattorn o~ contlguous haxagon~ with the maximum number of hex~gon-to-hexagon inter~ace3 for optimal reading and maximum information storage density i9 ref~rred to a~ a "honey-~0 comb patt~rn."
The contra3ting reflactive prop~rties utilizad to printth0 individual hexagon~ or cells o~ tho data array can be varied greatly within thc spirit o this lnvention. A~ ussd herein, "println~" m~ana deposlting materials having predatermined optical propertio~ on a ~ub~tr~e, or changing tho optlcal propert~es, as when "th~rmal" prlntinq is u~ed. "Printlng" also includ2~ the oml3sion to depo~it a matarlal having a pr~determined optical property on a portion of the sub~trate, where th2 ~ubstrate itself ha~ a distinct optical prop~rty. For axampla, ln printing hexa-gonal cell3 in black and whit~, if the sub~trate 1~ white, thenonly black cell3 mu3t ~ctually bo printod. Thu as used hersin, the white hexagonal cell~ ars al30 within the definition of the term "print" or "printed."
As u~ed herein, "op~ical propertie~" mean light ab-~orption, re~l2ction and/or refraction propertie~ o cells 13~9g~
printed in different media. Where cell3 ara printed in black (high den~ity black ink), gray (half tones of black) and white (no printing on a white substrate), as i9 the case in the pre-ferred embodiment o~ tha invention, the invention i~ said to have s three optical propertie~.
As used herein, and with reference to Fig. 1, "plural-ity of concentric rings" or "concentric ring3" 10 means two or more concentrlc rings 12, one of whlch L3 the interior area of a circular zona lS defined by the ~mallest radius "r" of the rlngs.
Fi~. 2 illustrates a portion of an electro-opti~a}ly ~cannabla label in ac~ordance with the principle of t~i inven-tion. A~ seen in Fig. 2, the label compris~s a multiplicity of adjacent printed hexagonally-shaped cell~, ~orm~d in a honeycomb pattern. Each of the individual hexagons i~ designated by numeral 20, and comprise~ 6 equal slde3 22. The interlor angle3 "a" of tho h~xagon are al80 egyal, each of 120 degrees. In the lllu#-trated embodlmont, the h~xagon ha~ a long vortlcal a~i~ y-y and a horizontal axi~ x~x. Th~ x-x dimen~lon of the hexayon 20 is some-what smaller than the y-y dimen~lon o the hexagGn 20 due to the geometry of a regular h~xagon.
In a pref~rred embodiment of the invention, a3 shown in Fig.- 3, utilizin~ a label 30 having dimenslon~ of approximately l" by l", ther~ will b~ app~oximately 388 hexagonq or cells 20 (taking lnto account the fact that, in th~ preferrad ~mbodiment, the conter of th~ label i9 occupied by an acguisition target 35 compriqed o a plurality of concentric rings~. ThesR contiguous hoxagons 20 naturally form horizontal row~ "R", defined by imagi-nary lines 31, and ve~ktcal column~ "C", defined by imaginary lin~# 33. In thi~ example a on~ inch by ono inch label has a total o~ 33 hor~ontal rows "R'~ and 30 v~rtical column~ "C" of hexagons 20. Eac~ individual hexagon has a "dlameter" of about o . a mm. Thor~ are mor~ roW# "R" than column "C" in a ~guare perimeter bounding a honeysomb pattern of hexagons, due ~o the geometrlc packing of the contigu~us hexag~n~.

~ 3~
Utilizing the hexagons illustrated in Flg. 2, it will be seen that the hexagons are aligned in staggered and overlapping vertical columns, with alternate vertically spaced hexagons having co-linear y-y axes. The y-y axeq of spaced hexagons 20 are in alignment with an exterior vertical side 2~ o an intermediate, diqplaced hexagon. The y-y axes of hexagons 20 are parallel to the two vertical bordar~ 32 and 34 of the label, as depicted in Eig. 3. Horizontal rows "R" are measured through the x-x axes at the mid-point of the hexagon 20.
A~ more fully described balow, the hexagon~ 20 are formed by a printing prosess which will print the hexayon~ 20 in two or more optical properties, for example, contrasting color~. Those colors may bo white 25, black 26 and also, optionally but prefer-ably, gray 27 a~ illustrated in Fig. 3, although other contrasting color~ may be utilized. It is po3sible to use only two contraqt-lng colors, ~uch as whlte 25, and black 26 a~ scen ln Fig. 2. In the pre~errod embodiment o the lnvention, three contrasting colors are utillzed, white 25, and black 26, aAd gray 27, illustrated in Fig. 3. Th~ particular ~had~c o white, black, and gray are s~lected to ach~e~ optimum contraqt for easo o identification by an eloctro-optical 3engor. The gray levol is select d so th~t it~ optical propertie~ ~all approximately e~ually botwsdn the ~.
optlcal prop~rtlas of ths white and black being used in creating the labol.
The label 30 of Eig. 3 may be ~ormed by using a discrete labol, having, in a preferred embodiment, a one squars inch area, or, i a~ acceptable color background is utilized ~praferably white), tho label may be printed directly on a package surface, without ragulring a discrste lab~l. B~causa o~ the importance of having a controlled optical property background or one o the contrastlng color~, it i~ pre~erabl~ to u3e a dl3crete Label, be~au~a the color o th~ label background is more ~aqily con-trolled~
The alignment of the hexagons ~rinted on the label in relatlon to the ~ides of the labal is importan~ or subsequently , ~ . I :. .

. ~ . , ,' ; ' . . ' ~ ~. ' :' ' ., : , ~ :

, .

~2~3 determining the major axis of the label aq deqcribed below. The label is printed ~o that the y-y axes of the hcxagon3 forming the ~ -honeycomb will be parallel to the vertiGaL side~ of ths label, 32 and 34, as shown in Eig. 3.
In "r~adlng" the hexagonal array, in order to decode the information contained i~ the lndlvidual~hexagon~, it i9 im-portant to hav~ a sharp color contrast between ad~acent hexagons.
For rea on3 described below, the fewer optical properties utllized to encode the hexagon~, the simpler may be the scanning equipment and software n~ce~sary to d~ode ths hexagonq. ~ow-evar, fewer optical propertie~ alRo decreaqe the data densi~y of thc labal. In a compromiqe betwe~n the amount of decoded infor-mation capable of b~ing ~tored on the label and the co~t o scan-ning multl-optical property label~, it has been found desirable to prlnt the encoded he~cagons with threa optical properties, namely tho colors black, gray and white. I~ the 3t~strate or lab~l ha~ a good white backyround, th~n white hexagon~ can be created by the absence of ink, and only black and gray hexagons actually need to be printed.
In th~ preferred e~bodiment oi the invention, the gray haxagonal c~113 aro cr~ated by printing the cell~ wlth blacX in~, , but only ev~ry ~ifth plxel of th~ pi~el grld o a dot matrix printer i~ ao printed in the illustrative example de w rlbed ~ -heraln. Thi~ i~ dona by the uqe of a half-toning algorithm, in a manner which i~ wall Xnown tn the art. Thi~ ailows a printer to print a predete~mined proportion o~ the pix~13 to define a given gray hexagon, whsrea~ a black hexa~on reguir~ printing every plx~l d~ining that hexa~on. The speciie half-t~ning algorithm u~ to print label of tha prefsrr~d embodimant i3 contained in the ~ourco code li3ting entltled 'i~A~E~" in the Microfiche Appen-dix, page 29, linas 39 to 48.
Th2 black hexagonal cell9 can be formed by printing with a standard black ink. A3 deqcribed below, ~he scanning an-aly~is so~ware of the deco~ing proces~ makas gro3s determina tion~ among black, gray and white re1ectivitie~, 30 that precise , ~ , .

" . . . .

9 ~ 3 color deinition is not necessary. On the other hand, if colors other than black, gray and white are used, or if various shade~
of gray are used, to create four or five color data arrays, the contra~t of ink shade~ must be much more car~fully controlled to S en~ure maa3urabla optical property differences among the various colors. It will be appreciated that the use of blacX ink is the simplest and ea~iest approach to creating a three optical prop-&rty honaycomb array of hexagonal cell~, and i~ the p~eferred cmbodimont o~ the inv~ntion.
Becaus~ of the ~quare ~h~pe of the label in the pre-ferred-embodimen~ and the nature of the hexagonal cell~, the edges of the honeycomb will contain lncompl~te hexagon~ 56; a~ seen in Flg. 3 the~e incomplete hexagons are not u~ed to convey any use-ful information.
In the pre~erred embodim~nt of the invention, the label also containe an acguisikion target. The acquisition target 35, ~e~n in Fig. 3, compri~ea a plurality of concantric rings of con-tra~ting color~ (3hown a3 black and white). The black rings are respectively de3igna~ed 42, 46 and 48, and the white rings are resp~ctively d~signated 44, SO and 52. The targ~t i~ preferably located in th~ g~om~tric center of ~he label, to make it less su~cepti~le to beinq damagad or de~troyed, in whole or in part, i~ thc p~riphery of the label is torn, soiled or damage~. Also, the size of the ~mago bu er tde~cribqd b~low), needed to ~tore thc data from the label befor~ thq labal target 1~ identi~ied, i~
minimizsd when th~ acquisltion target i~ in tho label center.
Tha numb~r of concontrlc rin~ uaQd in the acquisition targa~ may he varied, but t~e ~lx concontric ring~ 42, 44, 46, 48, 50 and 52 and thoir re~ulting int~r~acs~ as they vary rom white to black to white, etc., havo been found to be convenient and de~irable.
A pattern correlating technigue i~ u~ed to match a com-puted pattern of what the concentric ring~ are expected to be with the pattern being read. When the match occur~ the acquisi-tion target has been located as more fully deRcribed belo~. The , " ~

1329~3 .
speci~ic filter created and utili~ed in connection with the pre-erred ~mbodiment of the inuention may be found in the Microfiche Appendix, page 41, lines 51 to 52 pag~ 42, lines 1 to 8 and page 40, lines 19 to 41 under the file name "FIND.C."
s The acquisition target may b~ of any overall diameter smaller than the data array, to provide an area which may be 25%, and is preferably about 7%, o the area of the data array. Pre-~erably the acqul~ition target i8 ~izod a3 small a~ possible since tho area it occupios on the 1ab01 cannot carry encoded inormation.
In the pr~orr2d embodiment the diametor~ of the imprinted rings aro 3~1~cted 30 that the outside boundary of the external ring 52 is about 7.45 millimeters. Thu~, in Fig 3 the area of the acqui-sition target 35 occupies about 7% of th~ surface area of the one square inch label 30. In thl~ way, a ~atisfactory acguisition target 35 may be lmprin~ed on a one lnch square label 30 wlthout unduly interfering with the amount of lnformation which can be encoded in the he~agonal array ~hat ~urrounds the acquisition targl~t. As i9 th~ ca~e with th~ incomplote hexagons at the outer p~riphery o the lab~l 55, the fractional haxagon~ contiguous with tha outer boundary of the acguisition target 56 are not uti-lized for the purpvYo of encoding information. The width of each rlng i8 d~sirably about tha same a~ th~ ~ide-to-sida ~x-x AXiS i~.
Fig. 1) dimension of the hexagons, to acilitate re~olution. Si ring~ are convonlent. This is a reasonable numbor to facilitate location ~f the rings in a minimum label ar~a with a minimum of po~slblc ~also readlng~ from "3p~riou3" mark~ on tha labe} and oth~r "~puri~u3" mark~ not on th~ label, such as mark~ elsewhere on a conveyor bei~.
Th~ acgui~ition ~rget may take ~hapes other than con-centric ring~. For example, squar~s, spirals or he~a~on~ may be u~ed in order ~o crea~e tran~itions of contrasting concentric figures, so long as lineax sections through the acquisition tar-get will creat~ regular, predetermined and identi~iable color trangitiong, 8ugceptibl~ of being nsed by an electro-optical ~ensor and measurad by a suitable filter. It is to bo noted that, , "~", ~ . . :, ,, , ':

~2~9~3 although a spiral i not a collection of concentric circles, de-pending on the ~ize and radius of the'spiral, a cla~e approxima-tion of concentric circle9 can be achieved. A target of concen-tric ring3 i~ preferred, because the 3iqnal genera~ed by a scan through their center ha3 a freguency which is the same when sec-tion~ are taken in any direction throu~h the center of the con-centrlc rings. ~hi~ make3 identlflcation of the center qimpl0r, as more ully described below, and allow3 identlflcatlon of the location of the acquisition target with a one-dimen~ion ~earch of th~ analog or digital output of the s a~qr, although t~e proceqs of the invention may alternatively or ub3~quantly utilize a two-dim2nsional digital search for lncreased accuracy when a dlgital signal i5 boing analyzed.
As uqed hereln, "Concentrlc Ringsl' ia intended to em-brace ccmplote rlng~, partial ring~ in the fo~m of semi-circleq, qectors of concontric ring~ occupying bstwoen 180 degree~ and 360 degre~ and concentric spirals which approximate conGentric rings.
Since each hexagon may be encoded in three di ~erent optical prop~rties, in the preferred embodlm~nt, 1.585 "bits"
2~ of in~orma~io~ may be encoded in each hexagon (log 23) Obviously, if fawor or more optical propartia~ than three are u~ilized, th~ nu~bor of blts encoded in each hexagon will vary commonsuratoly. Tho oncodtng algorithm i~ ~truc~ure~ to achi0ve closa to maximum data den31ty and to increaqo the number of cell-2S to-cell optical property transitions, in order to facilitate the two-dimen~ional clock recovery process de~cribed below.
Flgur~ 4 illu~trate~ a 3 cell x 3 coll cluster of nine hexagonal c~lla 60, the ba~ic encoding unit utllizad in the prG~erred ~mbodiment of the invention. This i~ a deqirable en-coding approach, but i~ not es~ential. Other encoding u~its areea3ibl~, within tha purvlew o the inventton~ A more ~ully de~cribed below, the 3 cell x 3 c~ll clu ter of hexagon~ 60 are mapped to encod~ 13 bits of information if the cluster contains a ~ull co~plement o~ g hexagons; or Iess than 13 bits i~ the cluster is incompLete by having unusable hexagons. In a one inch square ~.:

132~
labal with a data array comprising about 88~ hexagons and an acquisition target occupying about 7 percent o~ the label area, about 1292 bit3 of information may be encoded.
In encoding each cluster, external, bottom hexagons 62 and 64 in each cluster 60, as ~een in Fig. 4, are limited in their respective optical properties, 90 that they are determined always to be different from intermediate and contiguous hexagon 66.
Thu~, only one bit per haxagon can be encoded in hexagon~ 62 and 64. ~n this way it i~ possiblo to encod~ 13 bits of information in cluster 60 by encoding 11 bit~ onto the remainlng seven hexa-gons. Since mapping 7 hexagon~ provide~ more possible combina-tion than are utilized (i_g~, 37=2187 combinations v~. 2'1=2048 combination~), 30me combination3 are rejected as, for example, all black, all gray, all white or sub~tantially all black, gray or white combinations. The reason~ for reguiring contrastlng colors o hexagon~ 62 and 64, compared to hexagon 66 are to guar-antea tran~ition3 neces~ary or the clock recovery 3tep and op-tional normalization proce3~ st~p described below and to assist in determining horizontal alignment of the data array, as described below. In ca~es wh~re encoding clusterq have 7 or 8 hexagons, 7 usabl~ hex?gons ar~ encoded with 11 bits and the eighth hexagon, if availabl~, is ~ncoded with 1 bit. Eor all oth~r partial clus ter~ 3 ~it~ are oncoded on ev~ry pair of hexagon~ and 1 blt onto oach remaining single hexagon as more fully describod below.
It will thereor~ be qe~n that the lab-l constitutes a particularly efficient, easy-to-read (by means of appropriate ~can~ing equ~pme~t and analytlcal sof~ware) label for encoding a very high denaity of inormation into a rslatively inexpensive, easy-to-print lab~l. A3 noted, tho prefelrod ombodiment utilizes a 33 row x 30 column packing of haxagons into a ono square-inch label, with an acqui-qitlon target repre~enting approximately 7%
of the total ~urface area of the lab~l. In practice, 13 bits of information are obtained from a clustsr of 9 hexagons, so that 1.44 bits of data are derived per cell. Thi is less than the : , ~32~9~3 thaoretical 1.585 bits per hexagon becau~e of the other con-straints of the encoding algorithm, since all 37 pattern~ are not u~sd, and some of the least optically desirable cell-to-cell transitions ar~ eliminated.
For rea~ons described below, in t~e preerred embodi-ment of the invention, it i8 desirable t:o incorporate a certain amount of error protection into the encoding of the label, so that the ac~ual amount of recoverabl~ information in the label is reduced in favor of a high degree of da1:a intagrity in the decod-ing proce~.
A~ may ~o readily appreciated by one ~kill~d in the art, the foregoing di~cussion of label embodiment3 employing hexa-gonal cell3 iY directly applicable to optically roadable labels utilizing other polygonal cell3. The disclosod mothod~ of "print-inq" optical propertie3 of hexagons apply equally to printing th0optical properties of other polygonal cells, whether in blacl~, white, gray ~through half-toning) or other colors. Similar con-straintq and advantage~ a~ to data density ~nure to labels printed with polygonal c~lls other than hexagon~ when the optical proper-t~o~ black and white, and optionally gray, ar~ utilized to printtho polygonal eolls. As with hexagon-containing labels, label pr~ted wlth oth~r polygonal encoding cslls may b~ "read" with ~canning egulpment of lesi complexity when only two optical prop ertie~ arQ utilized to encode inormation in the polygonal cells, in particular the colors black and white, becausa of the maximum contra t that i~ obtain~d with thase colors.
Infoxmation encoding procedures and the algorithm de-scrib~d or hexagon-containing labels are dlrectly applica~le to labol~ print~d with dif~oront polygon~l cells. Slmilar to hexagon-contalnlng labels, incompl~to polygonal cells which may appear at the border of th~ op~ically readable label~or that result from par~ial obliteration by the acguiYition target, compri ing a series of concentric ring~ are not used to encode information.
A "honeycomb patt~rn" compri~es an~ array of 35 contiguously-arranged hexagons 310, the geometric centers 311 of 13299~3 .~ .
which likewise lie at the vertices 311A of a "hexagonal grid" or "hexagonal array" 312, as shown in Fig. 11. Regular hexagons, i.e. hexagons having six equal sides and six equal interior an-gles, form hexagonal arrays that are liXewise regular in con-figuration, having three equally-~paced axe3 (Al, A2 and A3) that are spaced 60 degrees apart.
If the hexagons 320 of the label are irregular, but symmetrical, as for in~tance, if the hexagonq are stretched along two parallel sides 321, 322, the geometric centers 325 of adja-cent hexagons will describe an irregular hexagonal array 327, asshown in Eig. 12. Such an irregular hexagonal array will still have three axes (Al, A2 and A3), howev~r, the throe axeq will not be egually spaced i.~. the thres axe~ will not all be 60 degrees apart.
Although the hsxagonal array of Flg. 12 is not regular in nature, lt iq nevertheless a two-dlmonsional geometrlc grid or array having axes of predetermined qpacing. Thus, the locations and spaclngs of tho geom~tric cPnter3 of the hexagons located at the vertices o the inter~ecting ax~s of ~he hsxagonal array are also pr~determined. The gaometry of the hexagonal array ic then utillzod in th~ decodi~g procoss described b~low. Specifically, th~ filtration 3tep, performed on th~ tran~formed digital data corre~ponding to the image en~ed by the opti~al ~n~or, is adjusted to reflec~ tho predetermined 1 bel ~eom~try, so that the digital repr~sentation of the sen~ed label can be uqed to pr~cisely racon~truct the original grid. The recon~truction proces~ furth~r suppliaq th~ mi~sing polnts ~rom th~ hexagonal grid. The mi3~ing grid points occur because optical property tran~ition~ did not takc placa b~tween polygon3 of like optical proporti~s.
With irregular hexagonal grids of the type disclosed in Fig. 12 it will be desirable to adju3t the major axis deter-mination Jtep, ~tep (3)(e) of Fig. 7 o~ the decoding process done ater the Fouri~r transormation ~t~p of the proceqs to -~7 :~ ~ .... .

..

~ 1329~3 identify the major axis o~ the optically readabLe label. The ma~or axi~ of the label will hava tha ~eom~tric cent~r3 of poly-gonq lying along thi3 axi at different spacings than on the other two axes.
Lab~l configurations o the invention approximating the preferred embodiment containing hexagonal cells a~ described above are po~sible u~ing cer~ain po}ygo~al ceilq. Figura 13 il-lustrate~ a label conflguration utilizing polygonal cells 330 which substantially resemble hexagon~, hut which are 20-~ided polygons, rather than h~agon3. Similarly constructed polygons with more or les~ than 20 sides could al90 be printed. ~olygons 330 are partially contiguou~ unllX~ the imaginary contlguou~
h~xagonal G~119 331 in which they are depicted.
The interstitial space~ 332 o the Eig. 13 label em bodiment may or may not be printed With a differ~nt optical prop-erty than the encoded polygons. Intorstitial space~ do not carry ~ncoded information, therefore, thelr preqence lead~ to a lower data d~3ity ~or a glven optical resolution and per~ormance level.
Eurther, if the interstitiil space3 dispersed among polygons are of a diff~rent optlcal property than the adjacent polygons, more transition~ botween the optica~ propertias of the polygon9 and the in~or~titial spac~s could be qen~ed by the optical sensor an thu~ a hlgher clocX 31~nal energy would appear in th~ transform domain within the docoding proc~ss de~cribed in detail below.
Bocau3e the polygon~ of the Flg. 13 label are arranged on a hexagonal grld having three equally-~pac~d axc , the geomet-rlc center~ 333 of the polygonal cell~ 330 lio at th~ vertice~ of hexa~onal array 335. The ~pacing, location and spatial orientation of the c~nt~r~ o th~ polygo~ ar~ prod~termined and can be datact~d in thQ tran~form domaln of the decoding process.

.~ , .

.. .'', - ' ' ' ~ 3~9~3 The Fig. 13 labal utilize~ polygon~ substantially in the form of hexagons. Becauqe they so closely approximate a hexagon, the optical sensor at a moderate reqolution could "read"
them a hexagon~. The geometric centerq 333 of the polygons 330 do lie, however, at the vertices of the three equally-spaced axes (A1, A2 and A3) of th~ hexagonal array 335.
Elg. 14 illustrat~3 a similarly shaped (to polygon 330 in Fig. 13) polygonal figure 340 that ha3 been arranged to be totally contiguou~q. These polygons 340 can be approximated by an imaginary hexagon 341, as in Fig. 13, but no inter~itial spaces (33~ of Fig. 13) may bo found b~tweon the actual polygons. Such a contiguoua arrang~ment lq desirable to simplify the decoding proces9, but i3 not mandatory, ln the practice o the invention.
Polygons 340 are shown with their rosp0ctive gqometric centers 342 lying at the vertice~ of a hexagonal array 345. Again, as for the polygon~ 330 in Eig. 13, polygon~ 340 are 3ub~tantially in the shapo of he~agon~, and at a moderata optical reqolution would appo~r to be hexagons.
. Fiq. 15 i~ a blowup o a lab~1 as it would appear if pri~tod with a dot matrix printer printing 200 pix~18 per inch. ~.
Polygon3 360 of Fig. 15 illu~trate the shape of the geometric ~guro that wlll actually be printed in place of a hexagon with ~uch a dot matrix printer, beca~se o~ the pixcl danslty of tha printer. Print~rs wlth greater pixel den~itie~ qhould yie~d clo~r approximationq of a hexag4n than tha polygons 360 ~hown on Fig. 15. Thu~, polygons 340 o~ Fig. 14 and 360 o~ Fig. 15 are likely re~ulting ~hapes, due to th~ inhercnt limitationq of . 30 cortain printer~, of the printing proceqs for labels containing hexagonal cell_ or re ul~ ~rom deliberate effort~ to prin~ such polygonY

- ~ . . ... . . .

'' : ' " ........ : ~ , ' ~ .

-`` 13299~3 substantially in the form of hexagon3 in the flr~t instance. The shape of ~uch poly~on~ substantially in the shape o hexagon~
allow them to function, in a practical sense, as equivalents of contiguou3 hexagonal encoding cells.
As in the case of Fig. 3, the optically readable label o~ Fig. 15 al~o contains an acquisition target 370 comprising a qeri6s of concentric ring3 371 through 376. Like the hexagons on tha label o~ Fig. 3, the polygons 360 3ub~tantially in the form o hexagons in Fig. 15 are arranged in columns "C" and rowq "R,"
as bounded by imaginary line3 361 and 362 and 363 and 364, re~pectlvely. Alqo, as in the case of the hexagons of Eig. 3, the polygon~ o Fig 15 hav~ th~ir rospective geometric conters lying at the vertices o a haxa~onal array as defined by equally-zpacod axes Al, A2 and A3. ThU9, labels of the conig-15 uratlon ohown in ~ig. lS are readlly encoded and decoded in ac-cordancs wlth the processe~ di~clo3ed hereinbelow.
I f an alternative lab~l goometry is employsd such as utilizing a ~uare or rec~angular array, or the like, adjustments must b~ mad~ in tho two-dlmen ional cloc~ recovery process de-scr1bcd below. Th~ differ~nt geometry of the predetermined arrayrequires changes to bo made in the filt~r3 utili~ed in the filter-ing step o~ the two-dim~n~io~al clock recovery proce~q. The fil-tor~ oporate on the trans~orm0d dlgital data corr~ponding to th optical properti~3 of the polygons read by th~ sen~or ln the image domain. Such mlnor ad~ustment3 to the filtration scheme could aasily bo mad~ by a person of ordinary skill in th~ ~rt. In situ-atlons wher~ the predetermlned two-dim~n~ional array has unequally-~paced axe~, or i8 irrsgular ln configuration, it ~ay be desirable to identify th~ ma~or axl3 c~ th~ label prlor to performing the Fourisr tran~ormation o~ tha digital data repre~entin~ the opti-cally ~en~d image. Thi~ iY b~cau~e the gaom0tric centers of the polygon~ ara not egually spac~d along tha axe~.
Nonc~ntlguously-arrang~d polygon~ can alqo be utilized to create an optically readabls label ln accordance with the ' '' ~ ~' ' ' '.

" ~32~3 present invention. Fig. 16 illustrates a hexagonal array of quares 420, which are noncontiguously arranged with their respective gaometric centers 422 lying at the verticeY of a hexa-gonal array formed by the three Qqually-spaced ax~s Al, A2 and A3. It i8 apparent that the configuratl~n i~ a hexagonally-based configuration from the grid of imaginary hexagons 42} which can be overlaid upon the polygons 420, thereby forming interstitial ~pacas 425.
Similar array~ to the square 420 shown in Fig. 16 may ho con3tructed u3ing rectangle~. Fig. 17 illustrateq a multi-plt C1 ty of rectangle# 430 arrayed with the geometrlc centers of ad~acent rectangles lying at the vcrtices o a h~xagonal array formcd by intersacting axes A1, A2 and A3. Again, visualization of the hexagonal arrangemcnt i 9 aidod by the imaginary hexagons 431 in Flg. 17 overlaid upon the noncontlguous rectangles 430, thoreby creatlng interstltlal spacoa 435 between r~ctangles 430.
Flg. 18 llkewlso i11ustrat~s a noncontlguously-arranged label compri ing pentagons 440 having the geom~tric cent~rs 442 of ad~acent pentagons 440 lying along tho three egually-spaced axes Al, A2 and A3. Th~ geometry of the noncontiguous pantagons is more ea~ily vi3ual~zed by overlaying the pontagons 440 with im~ginary hexa~on 441, thu9 forming ~ntor titial spaco3 445 botwea~ pentagona 440. f Al~ernatlva h~xagonal array~ mAy be con~tructed where ~h8 axe~ of the array A1, A2 and A3 are equally spaced, but do not correspond to the axes of symm~try of tho polygonal figures them~lves. In~tead, th~ geometric centers of ad~acent polygons lie at ~h~ vortie~s o~ th~ inter3ecting axa~ o~ the array. Such an arrangomont i8 illustr~ted in Eig. 19, comprlsing a series of c4ntlguou~ roctanglo~ 450, having tho goom~tric contQrs 451 o~
ad~acent r~ctangle~ lying along axe~ A1, A2 and A3.
~igher order polygons may be 3imilarly arrayed on a predet~min~d two-dimensional grid. Fig. 20 show~ a ~eries of partially contiguou~ly-arranged octagons 460 dofining a mul~i plicity of interstitial space~ 461 among said octagons 460 The centers 462 of ad~acent octagon~ 460 are located at tho vertices :~ :

~ 329953 , . .~
o inter~ecting axes A1 and A2, thu~ forming an array of octagons 460, which may be u3@d in the practice of the invention. Inter-stitial 3paces 461 may be printed with an optical property differ-ent than iq used for octagon~ 460. However, thi~ is not mandatory in the practice of the invention, ~ince it is the location, ori-entation and intensity of the optical property at the center of tha octagonC 460, lying a~ a predetermi:ned position on the hexa-gonal array ormed by axe A1 and A2, thak i~ most important in the decodin~ proc0s It will be appreciated that althouqh a preferred embod-iment of the label has been diqclo~ed and described, many varia-tions of ths label are po~31ble without departlng from the ~pirit or scope of thi~ inventlon. For example, the label need not be one-lnch sguare. One ~guare lnch wa~ telected a~ a rea~onable 3ize of label, to achiev~ an acceptable data den3tty of 100 alphanumeric characters of information wlth a high degree of error 3~

:. ' ~ ~32~3 protection applied thereto, wlthout creating an excessively large size labe,l. It i~ desirable to have a one ~quare inch label, to reducs th~ paper and other costs associated with th0 printing, shipping and handling o such label~. Conventional bar code label3 of ~imilar sizo would have a radically decreased data den-sity. Using 4, 5 or more optical properties or colors to define ths hexagon~ will allow substantially more in~ormation to be packed into a given 3pace of hexagons o~ pre-datermined ~i e, but with a resulting increa3e in the complexity of the 30ftware and sen~itivity of ~he canning sy~em required in order to be able to recover that in40rmation. Thu3, for pra~tical purpose3, a throo optical p~operty, blacX, gray and`~hit~, ~ncodin~ 3ystem of optical propertiss i9 highly desirable. Al~o, the size3 of the hexagons and acqui~ltion target may bo varied widely within the splrit and scope of this invention.
Although "clu~tering" of h~xagons in 3 cell x 3 ~ell ~lusters ha~ b~en de~cribcd, other pattorn3 of clu~ter3 may be used or clu terlng may be omitted entirely and ths encoding algo-rithm may be directed ~pecifically to an individual hexagon pat-20 tern. Alao, tho relative amounts of encoded in~ormation davotedto the ~e~age a~ oppos~d to error correction may alqo be varied withln wid~ limits within the spir~t and scope of thi~ invention D~cribod balow i~ the ancoding proce~ of thiq inven-tlon, as applied to the preferred lab~l embodiment. It will ~e undor~tood that the pre4arred embodiment i8 being diqclo~ed and that num~rous comblnation~, variations and parmutations are fea-~ible within the purview of this invention.
Tho proce~s may begin with a predet~rmined series of da~a deslred to be encoded on a label. In a preferred embodiment, the label i3 a shippinq label, and the data is broken into two ~ield~, identl ied as a "high priority mesRage" and a "low pri-ority mes~age." It will ba under~tood, how~ver, that the inven-tion ls not re~tricted to two diferent message~ or levels of ,. . .

J

' .

~ .

1 3 2 ~ 3 ``~
priority. Many messages and levels of priority may be created within the quantitatlve limits o~ a label of given ~ize and num-ber of cellq.
For example, where the label i~ intended a~ a ~hipping label, the "high priority message" may con3titute nine characters, repre3enting the zip code of the recipient of the intended pack-age, parcel or letter. Nine digit~ i3 re}erred to because, al-though many individual~ and ¢ompanie~ ha.ve ~ive dlgit zip codes, nin~ digit zip codes are being uqed with increa3ing ~requency.
lQ Thersfore, ~n handling packages or d~llv~ry, th~ most important piece of informatio~ i the zip code. I~i3 determinos the general destination of the package and allow~ ~arlou~ scanning and pack-age control system~ to be u~ed to ~irect th~ package to the proper destination on truck~, aircraft, in a convayor ~y3tem and the like.
Tho low priority l~essaga may, or example, lnclude the name and ~hlpplng address, including zlp c~d0, of a reciplent of the lntendod package, a~ well a~ bllllng lnormation.
The re~on ~or creatin~ a high prlority me3~age and a low priority ma~ aga i~ to protect the high priority ma~sage with extra erro~ correction, to allow tha high priority mos3age to be placed ~encoded) in a mora central area of the label, where it is~ -leq~ liXely to b~ damag~d or destroy~d, and to permit the high prlority me~oage to be repeated and distribut0d in the low priority me~age 90 that, even if tha high priority me~ag~ i3 selectively d~troyed, thero i9 a hlgh pos3ibility that the high priority m~age ~an bo ro~rieved from the low priority me3~age. 3y locat-ing the hlgh prlor$ty me3sage in a central area, it may only be nece~ary to decode the high priority me3sage for ~om~ purposas, so that only a portion of the label n~eds ~o be processed, thus speedin~ up proces3ing tlme. Thi~ will occur, for example, when a parcel is on a conveyor and o~ly the zip code need~ to be deter-mined to ~ontrol whlch o~ several conveyor paths the parcel should take i~ the handling pro~ess.

~329~3 Because it iq of a lower priority, the low priority me~sage i~ not pre~ented twice on the label. However, as described below, both tho high priority and t~ low priority me~Yages may incorporate va~iou arror protection code3 and correction capabili-tie3, in order to maximi~e the likelihood that both messageq mayaccurately be retri~ved.
The u~e of error protecting characters as part of the encoded in~ormation can, in the preferred embodim~nt of this inven-tion, in combination wi~h an appropriate 3tored prosram and com-puter, cause the ~y#tem to correct an error durlng the decodingproce~, in the manner descri~ed below. The u-e o~ error protect-ing code~ ls well known in the art and is within the purview of the 3killed per~on in the art.
~n the practlce o~ the invention an operator creating a label may manually lnput the data to a ~uitable computer terminal whlch i~ designed, in the manner described below, to activate a printer to prlnt a label with the high prlority me~qage and the low priority m~s~age suitably encoded in the hexagon of the label.
It i~ not e3~ential to the invention that a hlgh priority message and a low pr~ority message b~ created, but it i~ da~irable in order to ~aximize the likelihood that the mo~t importan~ data to be encod~d will be retrieved. In the preferred embodiment the label i9 al o prin~ed with a centrally-located acgui~ition ~arge~
comprisiAg a plu~ality o~ concentric rings of two alt~rnatin~
contra~tlng colors, the color3 preferably b~ing two of the colorq utllizod to print the indivldual hexagons, and mo3t preferably black and w~ita to ensur~ maximum contra~t.
The operator manually inputting thi~ data will cause a -quitably prog~amm~d computer to encode each character o~ the input message and use ~ui~able field d~slgnator~, in order to crezte, in the operated computar, a binary bit stream, rapre enting the character~ of the me~age and ~uitably encod~d by ield to desig-nate ~h~ high priority and low priori~y mas~age~ and the relative position o~ each. Thiq operation is carried out by the program "TEXTIN.C" which may be found in the Microfiche Appendix, page 1, -35~

~', ' 1~2~3 lines 8 to 54; page 2, linej 1 to 54; and page 3, lines 1 to 36;
and is designated 110 on Fig. 9. A computer with the required features may be a Compag Deskpro 3a6 (with a 16-MHz clock and an Intel 80387 math coprocessor chlp).
Alternativel~, the proces3 may begin with the informa-tion to be encoded already contained in a binary bit stream, because, for example, it was received from a storage medium or otherwise created. Therefore, the message to be encoded can exist in a form ~hich i~ manually (with electronic a~qistance) converted to a binary bi~ q~ream or which begin~ aa a binary bit stream.
Once the binary bit stream ha~ been created or an error-protected bit ~tream ha been produced by the step9 di cu3sed more ully below, the bit stream mu~t b~ mapped in accordance with a predetermined mapping pattarn for the oncoding o the hexa-gon honeycomb of thia inv~ntion. Fig. 5 is a ~Icluster map" whichshows the individual hoxagonal ceLls of 3 cell x 3 cell clusters aligned in a grid or honeycomb containing 33 rows and 30 columns oi' hexagons. Each row i~ numbered, and each column lq numbered.
The row numbers range from 1 to 33, and the column~ range from 1 to 30. It can b~ ~oen that certain of the hexagons designated along th~ uppor ~urface and right-hand ~urface of the region map, and wlthin the geometrlc center o he grid ar~ dQsignated by X' a. Thl3 indlca~cs that tho~a hoxagon~ do not cont~in bit-mapped inormation. This i3 bocaus~ the oxto~ior X'~ reprecen~ partial hexagon- at th- edge of the label, thus causing each of theqe row~ to aach have one fewer hexagon. ~he int~rior hexagons desig-nated by X'~ represent sp~cos either occupied by the acguisition targat or incomplete hexagons around the perim~ter of the acquisi-tion target, 30 that those interior hexagor~ indicated by X's are not bit-mapp-d. All of tho hexago~s which a~ not ld~ntl~ied with X~ 9 are capable of recording lnformstion. In accordance with the preferred embodiment, each of the~e 3pace~ will be oc^
cupied by a ~lack (B), white (W) or gray (G) hexagon. As noted above, although various clustering and mapping technique~ can be utilized, the appllcation o~ thi~ invention may use clu~ters of 9 .

132~9 :` .
hexagonq in 3 rows of 3 hexagon_ each to deine specific bits of information, and, as also de~cribed above, 13 bits o information are desirably encoded ln each such 9-hexagon clu~ter.
In a data array comprising 33 rows and 30 column~ of contiguous hexagons, a grid of 11 row~ by 10 colu~ns of hexagon cluster~ each containing a 3 cell x 3 cell arrangement of con-tiguou~ hoxagon3, i9 formed and ~ay be ~risualized in connection with El~. 5. It will b~ appreciatsd ~ow~var that every row o 3 cell by ~ cell clusters withi~ the 11 cluster x 10 cluster grid wlll contain a cluster of either 7 or 8 hexagons becauq~ of the geom~tric paoklng o~ hexagons, and tha number will alternate from row to row. Thus, 6 clust~r3 containing 8 hexagon~ and 5 clus-ters containing 7 hexagons result from thi3 arrangement. Also, tho centrally located acQ,~lsition target create~ additional in-completo clu3t~rs. Flg. 5 thua provide~ a graphic representatlonof usable clu~ters of hexagons available for encoding with bits of in~ormation ~n a 33 row by 30 column data array o contisuous hexagon~.
With r~er~nce to Fig. 4, clustor with nine usable hexagons aro ~ncoded utilizing the following algorlthm:
Tak~ 31sv~n bit~ of information and map them i~to th~ set o seven haxagon~ identifled a~ a, b, c, d, ~, f and h.
~oxagons g and i ar~ u~ad to represent 1 bit each ln such a way aq to guarantee that each o~ them ls different from hexagon h.
Thus, thirteen bits o info~mation are ancoded in a complete 3 cell x 3 c311 cluster o~ nine contiguou~ hexagons.
~or partial clus~er~ o~ 7 or ~ u~abl~ hexasons:
Tako ~lev~n bit$ of information and map them into tho ~t of the ~ir3t s~ven u~abl~ hexa-gon-~.
The eighth hexagon, i~ available, is use~
~o repre~ent one bit.
For all other partial calls:

Map ~hree bit3 o informatlon into as many p~irs o hexagonq as pos3ibl0.

:. , ~ 32~9~3 ,, Any remaining sin~le hexagons are used to represent one bit.

Since mapping ev~n hexagon~ provideq more combination~ than eleven bits (i.e., 37 = 21a7 V3. 211 = 2048), some combinations of the hexagon ne~d to be re~ected. Th~ rejected combinations are cho~en to be those that provide the fewest number o transi-tions. To implement this, looX-up tab:Le~ were created to map the clusters in accordance with Fig. 5. The creation and use of the~e look-up tables i~ within the capabilitieq of a skilled programmer. Wit~ reerence to Fig. 9, tho program ~or creating the look-up table~ "BINHEX.LUT" 132 and "HEXBIN.LUT" 134 may be found in the Microfiche Appendix, paga 4, lines 3 to S2; page 5, lin~ 1 to 53; and pag~ 6, linos 1 to 34, and is id~ntified as "MK HEX ~UT" 130.
U~e o~ this blt allocatlon scheme allow~ 1292 bits of lnforma~lon to be encoded in a 33 row x 30 column data array of contiguou~ hexagon~.
The s~gu~nce in which the hlgh priority information and low priority lnformatlon is lo~ated throughout the cluqter map is p~edst~minod, dop~nding upon:
~a) Tha ~lza of the high priority me~sage;
(b) ~he size of the low priorlty message; and ~c~ The optimum location for the high priority messag in a protected place.
Utilizing the cluster map as illustrated in Fig. 5 as a î template, a ~tored mapplng program "MRMAPS.C" 140 operating on tho di~ltal data contalned in a ~torage m~dlum makes a predeter-ml~atioR ~ how to dl~tributa thd in~ormation -- both the high i priorlty me~sage and the low priority messag~ -- throughout the clu~ter map, as more fully described below. The mapping program 'i i~ iden~ified in th~ appended source code li~tings as "MKMAPS.C"'~ 140 and may be found in the Microfiche Appendix, pags 19, lines 3 , to 53; pago 20, lines 1 to 53; pa~e 21, lin~s 7 to 53; and page ', 22, line~ 1 to 42.
:j 35 , ., ' -38 ,: :

~ ~ 329~
In order to minlmixe the likelihood of error, and be able to correct errors, the preferred embodiment cf the invention desirably includes extensive error protoction and correction capa-bilities. For example, in a preferred embodiment having 1,292 S bit~ of information able to bs encoded in a one r~quare inch array of hexagon~ having 33 rows x 30 columns of hexagons, and an acqui-sition target occupying about 7% of the label area, it is desirable to utilize 36 hi~h priority message information bitR to encode a 9-digit zip code plus one~additional alphaAumerlc character, which may repre~ent a 3hipping code. In this example, it would also be desirable to uso 120 check blts for thr high priGrity me~lagr Thi3 is determined by the amount of error coxrection capability desired. Similarly, in th~ illu4trativ~ embodlment, 560 bits of low priority massage aro included; this includos 40 bit~ of high priority me3sage which 1~ incorporated in the low priority mes-sage. In the example, 576 low prlority mes~age check blts will be addsd in ord~r to maintain the security and faciLitate recov-ery of the low prtority me3sage. Thi~ ex~mple illustrate~ the much more lavish u~e of check bits in order to preserve and per-mit rocovory of ~he high priority message a~ opposed to the lowprlority m~s~ge. It i3 to be underctood that the foregoing in-formation is by w~y o~ example only and that the high priority mr3sag~ could be longar or ~horter, the low prlority mes4age longer or short~r, and the number of ch~ck blt3 greater or fower, dapending upo~ th6 particular application of the lnvention.
A "sy~tsmatic code" tako~ a sp~cific ma~age seguence and adds a di3tlnct error check s~quenco to th~ mr~s3age ser~uence.
A llnon-sy3tematic~ code tak~s a speci~ic me4sage ~equ~nce and inc~rporato3 ths rarror check seguence~with the mes~a~e sequenc~
90 that tho me~sage is no longor dl3tinc~, but i5, of course, rocover~ble. It is within the purview of ~his inv~ntion to use either sy3tematic or non-sy~tematic coding for error prot~ction.
The disclosure below i9 of a sy~tematic cod~.

:
~ ' :

: :.,. . ,, , -~32~3 ,., As defined herein, the step of "interposing error de-tection 9ymbols" includes systeMatic and~or non-sy~tematic coding system Various 3ystematic linear cyclic error protection codes are known in the art, or example, BCH code~, Reed Solomon codes and ~amming codes. In a preferred embodiment, Reed-Solomon code~
are separ~tely incorporated to protect the integrity of the high and low priority message~. Reed-Solomon codes ara very effi~ient and most u~eful when multl-bit character~ are being error-checXed.
Reed-Solomon codes are well known and it i3 to be understood that thi~ is simply a praferred embodiment, although many oth~r error correcting codes could be utilized in the invention. Reed-Solomon and other coding syRtems are diRcu.~3ed in, or example, Theory and Practic~ of ErrQr Control Code3, Richard E. Blahut, Addison Wesley, 1983, at paqas 174 and 175.
By way of example, some relevant in0rm3tlon about the Re~d-Solomon code ls set ~orth below. Speciflc characteristics of a Roed-Solomon cod0 can be ~pecified by the following param-eter~:

m - numbar of bits in each ~ymbol n = numbor of symbol~ in the block = 2m-1 k = numb~r o~ me~qage symbol3 (nu~ber o~
m~age bits ~ km) t = corr~ction capability in number o~
symbols = tn - k)/2 A nina-di~it zip code and single ~lphanumeric charac-tor ~or further ld~ntlfl~ation pu~pos~3 requir~s 36 b~t3 without error protaction in the example dascribed b~low. A Reed-Solomon : cod~ with ~h~ following parameters wa3 cho~en ~or the high priority mas~age.

m a 6 ~ 6 bit ~ym}:)ol9 ) n = 26 1 = 63 t = 10 Ther~fora, k = n - 2t = 43 Since only six 6-bit 3ymbols ara required ~o represent a 36-bi.t mes3age, the remaining 37 ~ymbol~ (43-6~ are padding ~ymbols, which are implied betwes~ the encoder and the decoder, , . ., ~ , , , ~3~9~3 ., and need not be ctored on the label. ThUc~ the total number of bits required on the label for the high priority mescage is (63 -37) x 6 or 156 bits.
This error coding 3cheme will be abla to correct a maxi-mum of up to 60 (10 x 6) bit errors, which amounts to 38.5% of the bit-c u~ed. Due to the lar~e number of implied padding sym-bols, the large error detection capability of thi3 Reed-Solomon encoding makes it extremely unlikely that the high priority mes-~ag~ will be read erroneously.
rh~ low priority m~ago wa~ encoded with a Reed-Solomon error proteetion code having difarent parameters, namaly:

m = ~ (R bit ~ymbols) n = 2a _ 1 = 255 t ~ 36 k = n - 2t = 183 Sinee thero ara 1292 bits available for encoding on the labol aecording to this illustration, a total of 1136 blt3 (1292 - 156 high priority messago bit~ and check bit~ are available for en-coding a~d check bit~ for the low priority mesYag~. Thu~, th~
remaining 904 bit (25.5 x 8 - 1136~ have to be implied padding bita. Thi9 allows 560.bits (183 x 8 - 904) for ~e inXormation con~ent of the low priority me3sage and 576 cheek bit~.
To f~rther ensur~ reeovory of th~ hlgh priority messag it i3 al30 ineIuded in th9 low priority mos~age. The Reed-Solomon error protoetlon eodo applied to the low priority mos~age permits oneoding of an additlonal 86 6-bit alphanumeric character~ and ha~ a maximum error eorreetion eapability of about 25. 4æ.
Utllizing ~h~ foregoing Reed~Solomo~ error protection encoding, th~ total numb~r of 1292 bita of i~ ormation available on the illu~trative labol ar~ di tributed ac follow~:
. 30 36 high priority in~ormation bits ; 120 high priority ehqcX bits 560 low priority informatlon bitq (includ~ng 40 bit~ of high priority me~cag~
ineorporated in the low priority me~age) ` 576 low priority cheeX bit~
.
Tha bit ~tream of data, ineluding the appropriate check bits for preserving the in~ormation, are assigned to individual : ~ :

132~
,.~ .
hexagons on the cluster map of Fig. 5. It wi~l b~ appreciated that a wide variety of distribution pattern can be utilized, recognizing that the i.mportant criteria to ba determined are:
(l) safer location of the high priority mes~age prox-S imate the acquisition target (if present on ths data array); and (2) ~reatlng a pattern which is reasonably easy toreas~emble when reading occurs.
The specific error coding proS~ram employed in th~
lustxativ~r example 13 contained in the Microfiche Appendix under the program "ERRCODE.C" at page 15, lin~s 1 to 52 and page 16, line~s 1 to 50.
Encoding for Reed-Solomon codas requires multiplication of tha me~sage code vector with a generator matrix. The matrix multiplicatlon is done u~lng Galols Fi~ld arithme~.ic. ~ddition lS of any two eloments o~ the 1el~ is obtain~d by parorming an exclu~lve "or" ope,ratlon between tho two elemant~. Multiplica-tion i~ performed via a "log" op~ration in the Galois Field. The log and antilog ars obtained by u~ing look up t~oleq generated from pr~me polynomials, specifically for tha high priority me~-sage: 1 ~ xq; and for th~ low priority mes~age: 1 ~ x~ + X3 ~
xa ~ x'. Wlth ~e~renco to Fig. 9; an auxiliary program "GF.C"
126 g~n~rat~s th~ look~up tables nece sary for tha Galois Field arithmetic. Auxiliary program "GE.C' may be ~ound within the ~icro~iche Appendix at page a, line~ 1 to 53 and page 9, lines l to 32. Th~ look-up tabler3 are computed and stored in the file "GF.LUT" 127 for u~e duxing encoding and decoding. The generator polynomlal g(x) or the ~eed-Solomon code is determined by the fo}lowing ~guation:

g~x) - (x ~ a)(x + a') ......... (x ~ a~t) wh~ro a i3 th~ primi~ivo ale~nt o the Galoi~ Fi~ld.
The generator matrix for the Reed-Solomon code i~ formed by per-forming a long division for each of the rows of ~he generator matrix. The,.kth row of the generator matrix is gi~en by the :"
., . . , . ~ , ' : ~ ' ' ', -1 3 ~ 3 remainder obtained from perorming a long division of xn k-i by g(x) The computation of the generator polynomials g(x) as well as the generator matriceq for both the high priority and low priority messag~s ls implem~nted accorcling to the auxiliary pro-gram "M~RSLUT.C" 125, which may be foùnd in the Microfiche Appen-dix, page 10, line~ 1 to 52; p~go 11, li~e3 l to 53; page 12, line 1 to 54; page 13, llnes 1 ~o 52; and page 14, lines 1 to 4.
The look-up tablo~ for the generator matrices are gonerated and ~o stored in the file "RS.LUT" 128.
In a preferred embodiment of the invention, labels con taining hexagons ar~ printed with ~tandard printing aguipment that i~ roadily available and inexpen3iv~. A printer having a 300 x 300 dot matrix per square inch capability will yield satis-factory results for printlng throe-color (black, gray, white) labols having 8a~3 haxagons plu~ a centrally-located acquiqltion ~ target. A printer with thss~ capabllities i3 the Hewlett Packardi Laser Jet Series II with O.5 ~egaby~es o memory and a 300 dot ', p~r inch graphic~ re olution. A 300 x 300 pixel grid having a 20 den~ity o~ 90,000 pixels p~r ~quare inch produces about 90 pixels por hoxagon ~n tho pref~rre~ embodiment. Each pix~l is a~signed a value of O or 1, reprs3snting a black or white pixel. This printer i~ utilized to print a two-color data array of b}ack or whita hoxagon3. It may al~o be used to print a three-color data array of black, whit~ and gray hexagons if a half-toning algorithm ~ utillzed to produca gray hexagons, a prevlously described.
? Roerring ~o Pig. 9, by meani o a stored program "MKMAPS.C," 1~0 a r~g~on~ look-up tabl~ "R3GION5.LUT" 141 of 34 row3 x 30 columns was created, which i3 analcgo~s to Fig. 5, but i 30 whieh was adapt~d to dc~ignit~ s~l~ctlon of black or whlte for the acqulsition target ring~. Individua} hexagons are cod~d for black, white or gray or as unu3able. ~ separate look-up table ~, "HEX MAP.LU~" 142 was created by a stored 3ubrouti~e of the pro-;1 ~ram "MKM~PS.C" which ~pecifie3 allegiance of each o the 300 x ~ 35 300 pixolq on the pixel grid to speciFic regionl in the :! .
~ -43-., . :
. , ~

~:

~329~3 ,,~ .
"REGIONS.LUT" 141, .i e., about 90 pixels per hexagon. Pixels be-longing to the inder rings are coded for elther ~lack or white.
Acquisition target ring~ are printed by ~irst g~n~rating a hex-agonal pattern on each region row then ~enQrating the ring Region~ partially or completely covered by the finder ring~ are rendered unusable in the "RSGIONS.LUT" 141. The foregoing pro-gram "MKMAPS.C" and subroutines may be ~ound in the appended sourca code in the Microfiche Appendix, pages 19 throuqh 22.
The error protection encoded bit qtream i~ mapped in accordance wi~h a predetermined seguQnco into the 11 x 10 cluster array ~f haxagon~. Still r~ferring to ~1g. 9, the 3eque~ce i~
specified by an order look-up table "ORDER.LUT" 151 generated by an auxiliary ~torsd program 2ntitled "ORDER.C", 150 which may be found in the Microficho Appendix, page 26, line3 1 to 47 and page 27, line~ 1 to 3. A ~tored program "PRLABEL.C" 160 and ~ound within the Microfiche Appendix at pag~ 17, lines l to 54 and page 18, lines 1 to 39, was utilized to as~ign values oE 0, 1, or 2 to the regions available for printing on the label, while leaving the re~ionq with a value o 3 unchanged. Gray lavel3 for each of the hexagons in a 3 c~ll by 3 cell cluster are a~signed in conjunc-tion with t~o stor~d program entitled "CELL CODE.CI' 170 found in the Microfich~ Appendix, page 23, line3 1 to 53; page 24, lin2s 1 to 53; and pag~ 25, line~ 1 to 43.
A pr~forence ~or storing the high priority mes32ge in an area proximate the acguisi~ion targ~t wher~ it will be le~s suscspt~le to label degr dation iY built into thl~ auxiliary ord~r prosram. Program "~AB~.C" 180 i9 ther~fore employed to ganarat- a bit ~tream suitable or input to tha la3er printer.
Prog~am "LABEL.C" 180 may be fonnd in the Microfich~ Appendix, page 28, line~ 1 to 53; page 29, lines 1-52; and page 30, lin2s 1-36.
It can be seen that the U9~ o~ black, gray and white permit~ a simple label printing proced~rs; because only black ink is neces~ary, when a standard hal~-toning algorithm i~ u~ed, in a ~' ' , . .

13~9~i3 manner which i~ well known in the art. If other color combina-tions are used (which is feasible), thQ necessity for printing in other colors obviously create3 sub3tantial complexit$es when com-pared with the three-color black-gray-white ~pproach or with a two-color blac~-white approach.
Thu3, when each pixel o~ the printer has been assigned a blacX or white valu~, the labal3 may be printed to crPate an encoded format, a3 illu3trated ~n Fig. 3, in which 30me hexagons are white 30me are gray and 30me are black, and in which an ac qui~ition target region, preferably of black and white concentric rings i~ ormed at the geometric center o~ the label.

~aving described how data is encoded ln tho label and 15 prlnted, lt i~ necec3ary to do~crlbe tho 3ubcequent label inter pr~tation or decoding proces3. It will be appreclated that it is desirable to pero~m the label lnterprstation function at very high #pe~d3, on the order of a fraction of a ~econd, in order to increa tha efficlency at which the package manipulation (or other manipulation or labol readinq) process take~ place.
Th~re ara two ba3ic approaches that can be taXen for capturing th~ ima~o in the label reading proc~s~. The label can b~ read at r~lativoly slow sp~ed9, u~ing a hand-hald static fixe focu~ scann~r: Alternatlv~ly, an electro-optical s~n30r, having a servo-controlled focu~ing ~echan1sm to permit dynamic scanning of rapidly movin~ package3 of variable sizes and helght3 is highly doslrabLa to achi~ve high ~peed operation. The decoding process and equipmont de$cribod bolow was demonstrated in connection with a fixed-focus ~canr.er. The proce~4 having the general capabili-tios described her~in with respect to a ctatic ~ixed-focus scanner is adaptabl~ to a dynamic 3canninq 3ystem with certain modifica-tion~ to the optical sy3tem a~ noted below. When manipulatinq pacXag~s~ at high speed3, it i9 desirable to have a high speed ~ca~ni~g mechanism which can r~ad labels travelling at a linear ~pe~d of about 100 inches per ~econd or more and pass;ing below a :
., .
1 : :
,, . . ' ~ , ~ .

~32~9~

fixed scanner location. The image processing function thus com-prises the following teps. Fig. 7 provides an outiine of the teps of tha decoding proces~.
1. Illumination of the ~abel When a package, parcel or letter is traveling on a high-speed conveyor, the area to be illuminated i~ quite large, be-cause the izes o ~he packages to be accommodated on the con-veyor could be guite large and variable. For exampl~, a 42 inch wide convoyor and pa~kag~ of only ~evoral inches in w~dth up to three eet in width (and ~imilar height~) are not uncommon in packag~ handling 8ygtem3. Therofo~e, the one squar~ inch label may be located anywhere acros3 the width o the conv~yor. Pack-ages are also lik~ly to be located at skewed angle~ with respect to the axis of movement of the conveyor belt. Th~ parcelj, pack-ages, letter~ or the like wlll have dlfforent height3, so that th~ label~ to be scann~d may bo located, for axample, one inch or le3~ abov~ the conv~yor, on the ono hand, or up to 36 inche or more above the conveyor, on the other hand, with re~pect to the maximum height packages that the descr~bed sys~em can accommodate.
In ord~r to properly illuminato the label~ in accord-anc~ with thi inventlon, eqpecially con~idering tha wide range of package widths, h~lghts and the angle of presentation of the i label~, it is de~irable to use a high-inten3ity light 30urce, whlch will r~fl~ct woll bas~d on the two or more optlcal proper-tie3 3el~ct~d for the labal. Ths light might be infrared, ultra-violet or visibl~ light, and th~ ht ~p~ctrum of uqable visible light may vary. Th~ t~chnlque or s~nsing ~he light preferably involv~3 sç~lsing light reflected rom th~ black, white and gray hexagon~ of the labol.
The illuminatlon ~ource must producs enough reflected light at the llght 3ensor ~for examplo a CCD device, a~ de3cribed bolow~ to permit the light sen~or to reliably distlnguish among black, gray and white or whatev~r optical propertlos of th~ hexa-gon are b~ing sensed. In a dynamlc qcanning sy3tsm an array of !a, ' ' . . ' ' ~ ' ' ' ............. " . ', .~ " ' .' '' ' ' ,. ': " ~ ' ' ' " ' "

132~
,, .
LED's could be u~ed to produce an illumination level of about 10 mW/cm~ in the label illumination area at the level of the label.
~he LED' 8 may be in an area array, without u ing a focusing lens, or a linear array, with a cylindrical focusing lens. A laser .light source, pa3sed through a suitable optical ~yRtem to provide a line source of illumination could also be uRed in the practice of thi~ invention.
The Jelection of t~a light source and the properties of tho light ~ource for the application in ques~ion are within the purview o th~ skillQd artisan. It i3 to be recalled that, since the label to be located iR only one square inch in maximllm di-mensio~, located at height~ of up to 36 inche~ on a 42 inch wide belt travelling a~ speed~ up to, for example, 100 linear inches per second, it is very important to be able to illuminate the labals properly in order to identify and locate the labels quite promptly.
In the caae of the static fixed-focus sensor utilized in the lllustrative oxample, an illumlnatlon level of about 2 milliwatt~/cm2 proved ~uitable for the practice of the in~ention.
Thi~ wa~ accompll~hed by means of a fluore~cent light ~ource.
2. O~lcal S~na~l~ o. the Reflec~ed ~abel Ima~e . Th2 ~oc~nd ~tep in the recognltion portion of the de- J
coding procs~s i3 to optically sen3e the illumlnated area with an electronically oporat~d ~ensor. $he camera/light ~on30r used in the illustratl~e example for a qtatic fi~d-focus 8canni~g system comprised a~ industrial guality color CCD telovlsion camera, such as model number WV-CD 130, available from Panasonic Industrial Company, One Pana onic Way, Secaucu~, New Jersey ~709~, fitted with a 50 mm 1.3 C-mount TV len~ includin~ a S m~ ~xtansion tubo, avallable from D.O. Indu~trieq, Inc. (Japan), 317 East C~estnut Street, Ea3~ Roche3ter, ~ew Yor~ 14445 and identified under the brand name NAV~TRONTM. The camera was coupled to an image captur~ board designated model number ~T-2B03-60, available from Data Tranqlation Inc., 100 Locke Drive, Marlboro, Massachu~
setts 01752.

.

132~9~3 Optical sen~ing may involve imaging the entire label, utilizing an area sensor such a~ the above-described camera and image capture bo~rd or, i~ the alternativa, may be accomplished with a linear array sen~or incorporatin~ a charge coupled device ("CCD") chip, wher2in the second dimension of the label scanning is performed by the movement of the package (and label). A suit-able CCD chip for this purpose i9 the ~lomson-C~F TEX 31510 CDZ, 4096 element high speed linear CCD image sen~or, available from Thomson-CSF, Divi~ion Tubes Elactroniques, 3a rue Vautheir B.P.
305 92102 Boulogne-Billancourt Cedex, F~ance.
For dynamic sy3tems involving the moveme~t of label-bearing packages on a co~veyor 8y8tem, it i~ de~irable to have a lony optical path between the label~ b~ing se~d a~d the light ~ensor. ThQ primary rea~on for cre~ting a long optical path is to reduco tho change ln apparent size or maqniflcatlon of the label aa sen~ed by a remote llght senaor.~ For example, if the optlcal path i3, say, four feet, thH lmage size for labels one inch abovs the conveyor wlll be very different from that for label~ three feat above the conveyor. If a long optical path is u3ed of, say, twenty feet, tho imaye size~ of tha 3ame labels are almo~t tho aa~. Thi~ allowa the area being ~ens~d, regardless of height, to flll all or ~ub~tantially all o the area of the llght 3en90r, ~0 achlev~ co~lst2ntly high lma0e resolution. I
an area sensor rather tha~ a line sen~or is used, the same prin-clple would al~o apply. Thi3 may be accompllshed by means of along optical path aJ deplcted in Fig. 6.
In order to be able to focu~ on labol~ o different height packag~s, a haight 3~n30r i~ need~d. An ultra~onic sensor may be used or a s~t of light b~a~s may ba brok~n by the package a~ a ~onso~. Either o the~ sy~tem~ 1~ u~able and may then actlvato a ~uitable adjustable ocusing mechanism with an open or closed loop mschani~m to ~ens~ and ad~u t the position of tbe optical sen~ing elements ( a., lenses and ~nsor) in rela~ion to each othar on a continuous basiq, as ~een in Fig. 6.
Fig. 6 iq a schematic view of a camera focusing and adjusting syst2m operable in accordance with the invention for .. .

: ` ~ ~ :, ,. ;

i~' ,: .. ' ' ' 1 3 ~
ad~usting the position of the camera light sensor in accordance with the helght of the package being sensed. Fig. 6 demonstrates a view of a suitable lenq 196, coil drive, hei~ht -en~or and feedback loop in accordance with the invention. In Fig. 6, the height sensor 206 may be an ultraqonic height sensor or a light beam which i3 broken by each packag~ travelling on the conveyor, for example. The helght sensor outpu,t is fed to microprocessor 204 which in turn actuate~ coil driver 202 to move coil 200 on which CCD 198 or other uitable light sen or is mounted. A shaft position sen~or 208 ~enses the posltion o~ coil 200 and its output to microprocessor 204 complete3 a eedback loop for ssnsinq and adjusting ths po~ition of coil 200.
Tho sen30r must b~ able to son~e the re1ected light coming from the illuminated label, and must also produce an analog ~ignal corra~ponding to th0 intensity of the re1ective properties of the labol a3 recorded by the indivldual pixels o the electro-optlcal sensor.
A ~uitable light sourco, as de~cribed abovo, may be mounted to a mounting surac~ above a conv~yor to bath~ an area extending acrcss the entira wid~h o the conveyor with a light o prodeterminod guality and inten~ity. The re~lected llght from th~ lab~l may be folded by a ~eries o reflectors and then is s~n3ed by an eloctro-optical en~o~.
The purpos~ of the folded optical path i9 to create a compact and therefore mor~ rigid ~y3tem.
Tho analog vldeo slgn~l output of the sen30r is then filtered. ~he analog el~ctrical signal is utilized ln con~unction wlth an analog b~ndpas~ fi~ter to det~t tho pres~nce of an ac-quisition target on tho data array. Th0 analog 3i~nal is then conv~r~ad to a d$gi~al slgnal using a conventional analog-to-digital conver~er incorporated in the ima~e capture board de-scribed b~low or by othor means known in tha art. In place o~ an analog bandpas3 11ter, it is pos~ible to 3ubstitute dlgital filtering clrcuitry to determine the prssence o~ the acquisition target by comparing the digital data representative thereo~ to ;
.

132~9~

the digitized signal output of the analog-to-digital converter, as more fully deqcribed ~elow.
An example of an area sensor having a CCD chip with a plurality of detectors and which ~as u~ed in accorda~ce with the 5 inVeAtion i3 the previously de~cribed P;inasonic WV-C~ 130 color CCD television camera. The analog signal output o the sensor wa~ communicated to the previously described Data Translation DT
2803-60 image capture board, including a 6 bit monochrome video A/D conVerYion for digitization and latl3r operations. 3y means of a suitable stored ~ubroutine the se~lenced digit l output of the image capture board wa3 saved in a m~mory device as an exact repllca of the image recorded by tho optical sensor.

3. ~roco~sina th~ Reflec~d Ima~e The mo~t important part of the invention ls to process th~ optically sensed ima~o in order to be able to recreate and orient with accuracy tho original label configuration and the color (optical properties) of each hexagon. This i3 done by using the following steps, after which th~ known pattern by which the lab~l was ox~ginally encoded and bit-mapp~d may be used to decode th~ information contained in the label.
(a) Locatina the Tar~e~ nt~

Prior to utllizing the above-d~scribed CCD television cam~ra and imag~ captura bo~rd, as outlined ln Fig. 10, an lnitialization prog~am "DTINIT.C" 250 was run to put th~ image captur~ board into a known ready sta~e and to loa~ tho output color look-up tables, followed by tho program "DTLIVE.C" 255 to put the ima~o capture board ln "live mode." The prsgram "DTGRAB.C" th~n causes the image capturo board to digi~ize ~he scene into a 240 row by 256 column image memory, with samples stored as 6 bit value~ right ju~tified in byte~. The foregoing program~ may be ~ound within th~ Micro~iche Appsndix respectively at page 31, llne~ 1 to 53; page 32, lincs l t~ 39; pag~ 33, lines l to 22; and page 34, lines 1 to 19. Two auxiliary programs ~ :

132~
"DTSAVE.C" and "DTLOAD.C" allow screen images to be transferred to and from a storage mediu~. Source code listi~gs for the fore-going progra~n~ may be found within the Microfiche Appendix, re-spectively, at page 35, line~ 12 to 33; and page 36, lines 13 to 33.
In ~irst acguiring the label image, a conventional analog band pas3 filter can be u~ed to identify two or more op-tical propertie~ of the acguisition target Concentric Rings.
These two opticaL properties are pre~erably the colors black and white becau3e the grsate~t contrast will craate the strongest ~lgnal enargy. In order to find a fixed patter~ o tran~ition from black to white to black, ~tc., it i9 da~irabls that a linear scan acros~ the acquisition target and passing th~ough the center o~ the target yield a uniform frequency response regarclless of label orlentation. Thu~, the target rlng~ ar~ opt:Lmally com-prlssd of contrasting Concentrlc Rings. ~le sensor output was then blfurcated and taken through two datection paths. One path detects all o th~ anargy in th0 output and the other mea~ures i the en~rgy at the ring roquency. When th~ two sutput~ are com-pared, tho ene~gy in the ring detector mo~t closely approximates the energy in the all energy detector when a scan through the acquisitlon target center i~ baing 3ensod. The acquisition targ~t center i2 located when this closest approximation occurs.
Source code ll~t~ng~ relating to th~ creation of a digital band-pass ilter and ~lterlng process may be ~ound in th~ MicroficheAppendix under tho Flle Name "EIND.C," pagos 39 through 43. How-ev~r, ln tha dynamic praf~rred embodimont of the invention, the - first filtering ~tap would pre~srably u~e an analog bandpass filter or el e a sa~npled analog bandpd~ filter, although a digltal ilter i9 U able.
; It is to be noted that tha acquisition targst locating step d~noted "EIND.C" 2ao ln Fig. 10 i~ indlcated as optional i~
Fig. 7, because a hand-held ~canner can be u~ed in the process o~
the inv~ntion~ in which avent the operator could properly place the ~canner to a~sure correct alignmsnt o the sen30r. This is, . .

. .

` . ,; , . ~ ,:. .
,. i . ~:

~3299~
o~ course, much slower than the use of an automated sensor and th~ use of the automated sensor i~ preferred in a h~gh speed operation. If an automated 3en30r (not hand held~ iq u~ed, }ocating the target i~ a required step of the proce~q.
As an alternative to an analog filtsr described above, a digltal bandpas3 filter may be con~tructed using the Park~
McClellan algorithm supplied with the software package "Digital Filter Designs Software or the IBM PC" (Taylor and Stouraitis, ~arcel Dekker, Inc., New York, N.Y., 1987~.
A one-dimen~ional digital band pasJ filter has been utilized i~ connection with the present invention to filter a normalized digital bit ~tream, as de~cribed below, through the following ~iltration ~ub rou~in~3. The band being fi-te~red is the expected ring ~requency. The one-dimensional digital, band-pass filtor was de31gn~d for a sampling rate of 400 pixels per inch and a le~gth of 125 plxHl~ (or 0.3125 inches~, and design~d to ba based upon tha sizo o the prlnted acqulsitlon target rings, as illu~tr~ated ln Flg. 3. The frequ~ncy was 300/16 Line pair~ per lnch, yielding a normallzed frequency, (where 400 line palrs psr inch = 1) o 300/16 x 400 or 0.046875. A filter with a pas~band ext~nding 5% ~elow this requency and 15% above was chosen bocaus~ label d~3tortionq typically re~ult in lmage 9hr~ nkaqe and thorefors an ln~rea~d ~raguency. Stop baAds ~rom~
15% bolow th~ f~cguency down to 0 and from 25% above the ring 2S frequ~ncy to 0.5 (Nyqulst limlt) were constructed. The filter coe~ficlant3 wer~ storQd ln the flLe "IMPULSE.LUT" 275, per Fig.
10, for lat~r operation~, delotlng the 1rst 62 coeficlents, becauso the iltor i~ ~ymmetrical. A flow chart 19 illustrated 1~ ~ig. 8. Eurther ref~renc~ may be made to the source code li~tings in the Microflche Appendix, under th~ file name "FIND.C", 280 st~rtlnq a~ pag~ 39.
A filtar of 25 pixol~ in length was constructed by sampllng the band pass filter at output interval~ corre~ponding to the measured horizontal magnification. Eor example, if the horizontaL magnlfication o the image i9 80 pixel~ pe.r inch, every . : :
.~ .

, . . ~ ~ .;
: . t 132~3 fifth sampl~ o tho filter would ~e u~ed (400/80 Y 5 pixels).
For non-integer steps, linear interpolation of adjacent filter samples is u~ed.
A ~econd 25 by 25 pixsl two-dimensional filter was also utilized. Sample valuss for this two-dimen~ional filter were based on the Euclidean distance of each point from the center of the filter, which wero scaled for appropriatQ horizontal and vortical magnification~. Linear intorpolation 1~ then u3ed fo~
-non-integer ~amplin~ intervals.
~he output of the abovo-mentioned one-dimen~ional fil-ter was sguared and smoothed with a ir3t ordor recursive lowpass filter, providing an exponential window o4 pa~t hi~tory. When the ~moothing filtar output exc~eded a predet~rmined threshold, an optional two-dimenslonal filt~ring st~p wa~ employed to verify the exiRtence of the target and to accurately determine its loca-tion, as de~cribsd below. The firat part of the two-dimensional filtering u3ed a roduced filter ~lzo of 10 pixel~ by 10 pixels to ~ave computatio~. Thl~ filtsr scans a rectangular area around the location det~cted by the one dimensional filter. If the maxi-mum two-dimenslonal corxelation exceeds a predetermined ~hreshold, then thc final 3tag~ of two dimen ional iltering, with the full 25 pixql by 25 pixel ~ilter, was appli~d to a small sguare window around tho maximum. If th~ b~st r~sult of this filter exceeded pr~determined threahold, tho cent~r wa~ det~ctod. If any of the thresholds were not exceed~d, the program partlally ~Idischarged~
the smoothing filtar and r~vsrted to one dimon~ional ~canning.
If one dimen~ional scanning completed without detecting the pres-ence o~ the acqui~ition target, the program exited with an error return. For any fuFth~r elaboration of the filtering process em-ployed in tho illustrative example, r~orenc~ should be made totho ~ource code listln~ in th~ Mlcrofiche Appondix, pages 39 through 42.

~L329~3 (b) Normali2ation of Sensed Ima~e Refle~ted light intensitiec recorded by the optical sensor empLoyed may vary due to variations in illumination, print den~ity, paper reflactivity, camera sensitivity and other reasons involving degradation to the lahel, for e~ample, folding, warping, etc. A~ an optional (but desirable) st~p, the reflected light sen~ed by the ~ensor and communicated to the memory may be nor-malized by conventional procedure Ucing techniques known in the art, a stored normalization program "NORM.C" 270, depicted on Fig. lQ, wa3 u ad to analyz~ ~he inten3ity level~ o reflected light rom the lab~l, a3 record~d by block3 o~ pix~ls in the 3canner, to ~ind the minimu~ and maximum reflacted light intensi-tie~ recordod for the data array. The ~equenced digital output of the abov~-described scann~r and lmage capture board combina-tion wa~ loaded rom memory to the computer to be urth~r operated upon by said stored normalizatlon program.
Utillzlnq an equation y = mx ~ b, where the minimum in-ten ity sub~tituted for x will yield a v,alue of y = 0 and a max-~mum i~tensity ~ub~tit~ted for x will yield a value of y = 63, 20the r~cord-d int~nsities o re1ected light for aach pixel were ad~u~ted ~o th~ tho blacXe~t black and tho whlte~t white present in th~ ~torod imag~ wore o~tablished a3 th~ ~tandard, and the other shades of blacX, white and gray were ad~ust~d to thoce standard~. The normalizatio~ step thu3 make~ the sensed image eacier to proce3s. Normalization was carried out using the 3tored program "NORM.C" found in the Microfiche Appendix at page 37, line~ 10 to 52 and page 38, line~ 1 to ll. It will be appr~clated that othsr, moro 30phlsticatod normalization proce-dur~s known ln the art may bo appllsd.
(c) escalin~_~he Ima~e t For sub~eguent computation3, the 3tored replicated label ima~e is re~caled to create an image with equal horizontal and I

, .:
~ .
:

132~9~ 3 vartical magnification. Again, this is a~ optional step, but it facilitates the ~ast and accurate recovery of the encoded informa-tion. The rescaling operation was performed to give the image a uniform hbrizontal and vertical sampling resolution of, for ex-ample, 150 pixelJ per inch, a~ used in the illustrative static ixed focus ~mbodiment o tha inventlon.
Tha re caling oparation occurs by computing the frac-tional row and column addre~ses of sample~ at 1/150 inch, based upon the known hori~ontal and vertical ~lagnification. Each point on the new unlform rescaled image is th~n extracted from an ap-propriate qet o polnts on the image replicated in the storage mediu~. Bilinear interpolation is used to approximate tha value of point~ at fractional addres~e~. The re~caling plac0~ the center of the lab~l at a known position in memory. The rescaled image ls stored for later use in the searching stap. All subse-guent process steps then a~sume that a rescaled lab~l image is centerod on a known po~ition on the grid, but it should be noted that this doos not indlcate tho orientation o~ tha label, which may still be skewed wlth respoct to the s~nsor. The rescaling operatlon ia carriod out undor the control of a stored subroutine fou~d 1~ th~ ~ource coda li3tlngs withln tha Microfiche Appendix at pago 42, lino~ 14 to 52 and pag~ 43, line~ 1 to 14. ~;~
(d) Two~21m~n~10nal ~lock R~over~
Tha ~ext seguence of steps of the process are referred to collectiYely as "two-dimonsional clock recovery." The steps ara performsd by a ~ultlble stored program and ~ub~outines en-titlod "C~OC~.C" 290, d~plcted on Eig. 10, and ound within the ~icro~ich~ Appendix at pagos 44 through 51. This oporation is per~ormed in two dimension~ on the ra~caled lmage to determine przcisoly the positlon o each hexagon on tho origlnal data array.
Tho purpose of clock recovery i~ to detarmine ~he sampling loca-tions and to correct ~or the efect~ of warping, curling or tilting o the label, since the label may not be perfectly flat.
This ij an important part of the process and its application is .

1329~5~
not limited to hexagonal encoded labels. It may be applied to other proces~e~ for decoding an encoded label comprising a regu-lar, two-dimensional grld, such as squarés, triangle , etc.
One-dimen~ional clocX recovery is a generRl concept which i~ well understood in the signaL proces3ing field. Two dimensional clock recovery i5 an extenslon o that process and will be understood, upon reflection, by the skilled technician.
It will be understood that the "clock recov~ry" term is somewhat confusing to the non-export, since it does nst relate to timing.

(i) Edqe Enhan~n~_and Non-~:Lnear O~aration Th~ first 9tep in accompli3hing clock racovery may be performed by variou~ non-Linear mapping op~ration~ known in the art to create siynal components at a specified clock frequency that are missing from the digitiæed ima~e output from the optical sensor and image capture board. The purpose o~ non-linear map-ping i~ to take tho (preferably) normalized and rescaled image which exist~ at thig po~nt in the proce3s and to form it into a two-dimensional non linear map whlch enhances the tran~ltions be-tw~n ad~a~o~t contrasting hexagon In tha preferred embodimentof the pro~ent invention, tXis is done by tandard deviation map-ping. Thi~ t~p can also be performed by filterlng with an imag ~.
differencing kernel, sevexal mean3 for which are known in the art, such as LaPlace or Sobel kornels, and then an absolute value i9 determined or squaring of the results is performed. These procedurac may b~ fo~nd in the text Digital Imags Processing, Rafael G. Gonzalez and Paul Wint~, ~ddi~on W&~l~y, 1977.
In ~tandard deviation mapping, th~ image wlth undif-fere~tiated cell-to-cell edge3 i9 stored in memory. ~ ~tandard .dsviation map i then crea~ed to locate th~ edges of ad~acent contra~tlng hexagons by determining the standard devia~ions of 3 x 3 groups of pixels (this i dif~rent rom the 3 cell x 3 cell clusters), to determina the standard deviatlon3 of the pixel in~ensitias. The ~tandard deviatlon computation are perormed 3; to determine the regions of pixels having a ~lxed color (the , . : :.

,,~ ~ . .

~32~9~3 lowest ~tandard deviations), reprasenting the interior of a hexa-gon or the interface between two like-colored hexagon~, as op-posed to ~he groups of pixel~ having higher ~andard deviations, which represent tranqitions from a hexagon of one color to an ad~acent hexagon of a contraqting color. Because adjacent hexa-gons frequently hava the same color, the standard deviation map will not completely outline every hexagon. Missing border~ or edges between hexagons wlll result from the fact that the stan-dard d~viation mapping process cannot distinguish interfaces between hexagon~ of the same color. Further aqpect~ of the clock reGovery proGes~ ara intended to regenerate theJe missin~ transi-tion~.
Tha decoding process of the instant invention may be utilized for any of the previously described label embodiments, as illu~trated in the accompanying figures. Encoding units of variou3 geometries may easily be accommodated and ~uch optically encoded polygonal cell~ may be array~d with the geometric centers o~ ad~aaent polygonal cells lying at the vortices o~ a known, pred~termin~d two-dime~sional array.
Whon the optically readable label~ of the instant in-ventio~ are "read" with optical sen~ors o the type~ described hereln, the partlcular g~om~try or shape of tho individual en-coding units or poly~onal c~lls is not determined by the optical sen30r. Instead, the ~ensor simply samplas the optically read-2S able label at a known number of samples per inch and records the intensity o the reflected light corresponding to the optical property of ~h~ particular ~a~ple area that ha3 h~en imaged.
The~e values are then ~torod in a storago modium for later proc~ing. In other words, the ~lectro~optical sensor records tho avorago light inten~ity sample area-by-sample area across the entire label ~ur~ace, regardless of whether or not anything is printed on the label. Thi~ is wha~ i~ meant by ~oring the ima~e with undifferentiated cell-to-cell edges in memory. Eor thls rea~on thc decoding proce~s is readily adap~able to reading optically readabLe labels o~ widely varying configurations, so , ., . .. ,.;. .

, , . ' ,:
- ,,, , , ~

1~299~3 long as the centers of the polygonal encoding units lie at a predetermined spacing anfl direction on a two-dimen3ional array.

In practice it ha~ been found that alterations of the hexagonal encoding cell-based syqtem, a~ in the case of label embodiment3 employing polygon~ substantially in the shape of hexa-gon~ as illustrated in Fig. 15, rasult :Ln negligible reduction of the system's performance. Utilizing polygonal shape~ with poorer pacXing characteristics, or array3 o4 partially conti~uous or lU noncontiguous polygon~ rather than contiguou~ packin~, will then re ult in a poorer, but nevertheless, usable syqtem performance or many applications. At ~ome point, however, ~ue to the opti-cally unresolvable high frequency components of lower order poly-gonal encoding cells, inefflciont cell packing and predetermined lS two-dimensional array~ ro~ulting in largo interstltial spaces b~twe~n polygon~, the ~ystem p~rformance will ~all to an unaccept-ably low information ~torage and retrloval capacity.
Tho acceptability of the 3ystem depend3 on the quality of the ~ignal recover~d by tho electro-optical ~en~or. By alter-ing the ~nslng ~y~tem, a~ for oxampl2 by increa~ing the number :: . . .
~:

. -. -.

132~9~3 of qample~ per unit area across the label surface, one could im-prove the signal recorded by the sen30r and improve the information storage and retrieval characteriqtic~ of such partially contiguou3 and noncontiguou~ label configurations.
Such adjuqtment~, in order to make such les~ desirable label configurations usable, would be within the abilitles of one of ordinary skill in the art of electro-op~ics.
The process, therefore, allows a wide variability in terms o the label articla, optical ~i~nal acguisition means and Rignal proce~sing. Polygonal cell~, of either regular or irreg-ular form may be used as encoding units on the optically readable label~ of the lnven~ion. Fur~her, so long a~ the spacing and direction of the center~ of the poly~on~ ar~ known in relation to adjacent polygonal cell3, the polygonal encoding cells may lie on a predetermined array, other than a hexagonal array, and the poly-gon~ may be arranyed contiguously, partlally conticJuou~ly or even noncontiguou31y on the optically readabls label.
As explained in greater detail below, the nonlinear mapping technique~, sp~cifically tha standard deviation mapping technigue~ di3clo~ed hsrein in relation to the preferred embodi-ment, facllltate rocon~truction of the mi ~ing transition~ or ~dge~ betwo0n polygonal coll~ of like optical propertie3. More-ov~, tho oam~ foaturo may ovqrcom- ~h~ lack o~ t~an~ition3 between polygon~ and intcr~titial ~pace~ between polygon~ of like opti~al propertie~. Thi~ i~ the situation whon lab~l configura-tions contalnln~ partially contiguous or noncontiguous polygons aro utilized. Thl~ featur2 i9 a~compli3h~d throu~h the following Fast Fourier Tra~fo~matlon, iltering and i~ver~e Fajt Eourier Tran~ormation step~.
An optlonal techni~ue utllized in tha preferr~d embod-lment of the p~s~nt invention reduc2s the computatlons needed to generate the ~tandard deviatlon map. Normally, to compute the 3um of the nine pixels in each-3 x 3 pixel block, eight addition op~rations wo~ld be needed. Thi3 may be cut in hal~ by replacing each pixel of the imag~ with the ~um o~ itself and the pixels :
: , , ~ , . ,:
- :

.

~32~3 immediately to it3 left and right. This requires two additions per pi~el. T~hen, the 3ame operation i~ performed on the new image, except the ~um i9 computed for pixel3 immediately above and below. This requireq two more additions for a total of four.
It can be 3hown that, at the end of these steps, each pixel has been replaced by the sum of itself and it~ eight i~ediate neighbor~.
Standard deviation mapping is the desired technique for creating this map of hsxagon3 corresponding to the original data array but with mi33in~ tran~itions between origlnal he~agonY of the ~ame color. The 3pecific standard deviation mapping tech-nigues utilized in conjunction with the illustrative embodiment may ba found within the source code li~tings in the Microfiche Appendi~ at pago ~5, line~ 14 to 53 and page 46, line~ 1 to 4.

i~dowlna ~ he next ~ubroutine, called wlndowing, is optional.
Windowing wa~ used in the practice of the invention to reduce the i~tensity of border~ which are not a8~0ct ated with hexagon outlines.
Thege border~ o~cur at two locations: the target ring3 and the uncontrolled ima~e ~urroundiny the label, A weighting function i9 utilized to reduce the inten3ity of these area~. The details of how to use wlndowing a~ a precur~or to a Fa~t Fourier Transfo i3 within the purview of the 3killed artisan. The windowing pro-cedure utilized may be found within the ~ource codo li~tings con-tained in ths Micro4iche Appendi~ at page 46, line~ 6 to 22.
(iii) wo-D~men~io~al Ea~t_~ou~su~an~orma~ion A two-dimenslonal Fa~t Fourier T~an~formation of the digltal valuo~ corrs3ponding to tho.~optionally) wlndowed ~tan-dard deviation map i then performed undar the control of a commercially-available 3tored program. In operation, a computer performq a Fa3t Fourier Tran orm of the image generated by the prior ~tep to yield a two-dimensional repre3entation of th~ spac-ing, direction and inten~ity o~ the interface3 of contrasting ~-60-, :

~ 3 ~ 3 hexagon~ identified in the standard deviation mapping step.
Simply stated, the East Fourier Transform is a measure o the spacing, direction and intensity of the edge~ between h xagons, where known. Thus, the regular spacing and directionality of the - 5 hexagon boundaries will cause certain points in the transform domain to have a high energy lav01. The brightast point will be at 0,0 in the Transform plane corresponding to the DC component in the image. The six points ~urrounding the central point repre-~ent tha ~pacing, direction and intensity of the edges between hexagons.
It will be recognized by one ~killed in the art that, as for hexagons, a two-dimensional repre~entation of the spacing, directlon and intensity of the interfac2~ of contrasting polygons identified ln the preceding standard deviation mapping ~tep can al30 be computed by performing a Fa t Fourier Tran~form of the disital data corresponding to the non-iinaarly mapped sensed Label image. ~hu~, the spacing and dlrectionality of the polygon bor-dar~ will caus~ cortain point3 in the tran~form domain to have high energy. The number of points of higher energy surrounding the csnt~r point at the 0,0 coordinate of the tranqform plane will depend on ~he goometry of the particular polygonal encoding cell u~ed to ma~e tho optically readable labal. As for hexagons, however, ~ucA polnt3 surrounding the central point wlll represen~
the ~pacing, direc~ion and intenqity of the edge~ between polygons or the edge~ between polygons and interstitial spaces if the label configuratlon is either partially contlguou~ or noncontiguous in natur~.
Sinco the imago i~ real (not complex) valuod, the Tran~form domain i~ polnt sym~etric about the origin. Thus, only a half plan~ o the transform domain must be computad, thereby saving nearly a factor o two in computation time. Elimination of these computation~ also reduce~ the amount of effort required in the sub eguent imag~ tltering and Inverse Fast Eourier Trans-formation teps. Tha FaYt Fourier Transf~rm program utilized in connection with the illuqtrative embodiment of a tatic fixed ~, . . . . .

:

~329953 ocus system was the commercially-available subroutine R2DFFT
from the 87 FFT-2 package from Microway, Inc. of ~ing~ton, Ma3sachusott~.

(iv~ Filterlnq thelm~e A filtering process i8 now required to reconstruct the complete outline of all of the hexagons in the image domain, utilizing th~ transformed digital data. This is done by elimi-natlng any tran~form domain point3 that do not correspond to the dosired spacing and dlroction of hoxagon boundarie3 identiied in the ~tandard deviation mapplng step. Six prominent points in the tran~form domain arise because o~ the hexagonal honeycomb struc-ture of the label. Only thre~ points in the transform domain are actually id~nti~ied, because the imaga is point symmatric about ths origln, and the second three points may be inferrecl from the irst three. In the preferred embodiment, flltering i performed in threo steps to eliminate transltion~ rom the standard devia-tlon mapp~ng 3tep, which are too far apart, too clo e together, and/or in tho wrong dir2ctlon.
Fir~t, high pa~s filt~ring i~ porormed by zeroing all points withln a pred~termined circla around the origin of the ~;
Transform domain, but at a distance oxtending outward from the origin, short of the 9iX promin~nt point~ arrayed ln the shape o~
a hexagon, in tha graphic transorm domain. Theqa points corres-pond to spacing3 greatar than the hexagon 9p cings and thu~ carry in~ormatlon portainlng to the mi~3ing tran3itions in the label imag~. To rocreato missing tran~itlon~ in tho lab~l image, it is ~ece~sary to oliminat~ the information about the ~iasing transi-tion~ i~ the Eourier Tran~orm domain.
Next, all points ou~side a cortain r~diu~ beyond the ~ix promino~t pointa in th0 Tra~orm domain ar~ zeroed. These correspond to spuriou3 tra~sitions that are ~paced too close to gether. This operation combines with the first on~ to form a ring of remaining points. Cr&ating this ring i9 ~quivalen~ to performing spatial bandpass filtering. The inner and outer radii , : :- - -:

13~9~5~ ~
o~ the annulus are determined by the expected ~pacing of the hexa-gon outlines. Since the hexagon "diameter" i9 expected to be 5 pixels i~ the exampla being described; and for a transform length of 256 pixel~, the hexagonal vertlces in the Tran~form domain should be 256/5 - 51.2 pixels away from the center. Accordingly, a ring with an inner radius of 45 pixels and an outer radius of 80 pixels corresponds to hexagon diameters of 3.2 to 5.69 pixels was u3ed. A filter with a pre~erence for pa~3ing higher frequen-cie~ wa~ uqed becau3e label deformation~, such a~ warping and tilting, cause image shrinkaga.
A~ter parformlng the qpatlal bandpa~s iltering des-cribed above, an annulus with 3iX prominent pointY exists, each point having equal angular ~p~cing with respQct to the center (0,0 polnt) o the transform domain. To complete the task of re-jecting und~ired information in the Tran~form tomain, a direc-tional filtering step is employed. Any point at too great an angular di~tance from the prominent reglon~ in the Transorm domaln i~ zorood. Thi3 has the eff~ct, ln the ima~ domai~, of removing any edge~ that do not occur in one of the three direc-tio~s dictated by the hexagonal honeycomb tiling pat~ern.
To per40rm directional filtering it iq nec~ssary to~ind the mo~t prominent point r~maining after spatial bandpass filtering. Thi8 point i9 a~sumed to be one of the six prominent~
points o~ the transform dom~in re~embling the vertices of a hex-agon. Elvo oth~r prominent polnt~ at the ~ame radius from thecenter and with angular ~paciny of multiple~ of 60 degreeq are al~o evldent in the transform domain. Therefore, all other point~ with an angular di3tance o~ gr~ater than 10 degrees from any of these polnt~ are eliminated. Six wedgo~ of the rin~ re-main. By this directional ~iltering ~tep, any information ofincorrect spacing or directlon in the image domain is eliminated.
Eliminatlon of thi incorrectly spaced information enaoleq the r0storation oi a comp}ete Gutlin0 of each hexagon in the image domain.

: : :: . .

, 9 5 ~
.
The or~going iltering steps are performed under the control of storad qubroutines contained in th~ source code list-in~ within the Microiche Appendix at page 4~, lines 26 to 53;
pag~ 47, lines l to 52; page 48, line~ 1 to 52; and pag~ 49, S lines 1 to 46.
The foregoing discusslon of the iltration scheme em-ployed or the preferred la~el embodiment comprislng contiguously-arrangsd hexagons requires modification ~hen different predeter-mined two-dimensional arrays are utilize~d for the optically readabla label. It will, nevertheless, bs appreciated by one ~killed in the art that only slight modificationY to the filtra-tion schems are reguired to accommodat~ the different label con-~iguration3 that hav~ been previously described herein, and which are illu~trated in ehe accompanying drawings.
lS Onco the individual polygonal encoding cells are de-clded upon, lt i~ predatermined that their respectivo boundarle3 wlll have cartaln angular spacings, and a givcn number of ~ides of given l~ngth. Next, it is nece3sary to d~termine the rela-tion3hip of ad~acent polygons, as ~or instance, whether they will be contiguous, partially contiguou~ or noncontiguou~. Also, the geometric array upon which the geometric center~ of the polygons ~ill ba arrang~d n~eds to be e3tablished. Sinc0 the foregoing lab~l qeometry iY prodotermlned a p~r30n of ordlnary sXill in the art can con~truce tho approprlato filtratlon ochem~ to fllter the energy point3 in th~ transform domain, ~o that only the brightest point3 corre3ponding to t~e appropriat~ spaclng and direction of polygons boundarie3 i~ oparated upon by the in~ar~o Fast Fourier Transform ~ubroutlne.
Concerning the actual filters constructed, t will be approclat~d that it is noces~ary to con3truct an appropriately d~men~ionod ~patlal bandpa~q flltor ba30d upon th~ predetermined spacing of the polygonal encoding c0113. Then, it is desirable to con~truct a directional fiLter to filter out energy points .~

:

~ ~329~53 .
other than the most prominent points corresponding to the axes of the predetermined two-dimensional array of the polygonal encoding call9. Thi3 eliminates any information concerning the incorrect spacing or direction of the polygonal encodin~ cells in the image S domain and the interstitial spaces if present. By aliminating such incorrect information a complete array of the centers of the polygonal encoding cells can be recon~tructed in the image domain by means of inVoX90 Fa~t Fouri~r Tran~o~mat10n in accordance with ths process 3tep described below.
(v) To actually raturn to the image domain, thereby restor-ing the outline image of the contiguous hoxagon~ o~ the data array, it is d2sirable to perform a two-dimensional Inverse Fast Fourier Tran~form (2D-IEFT) on the ~iltered tran~for~ domain data. The inverqo tran~form is implemented by a standard two-dimensional Inverse Fourier Trans~orm subroutine (R2DIET) available in the 87EFT-2 package from Microway, Inc. of ~ing3ton, ~a~achusetts.
Upon co~pletion of the invar~ Transform ~tep, the outline of evsry hexa~on i~ raotored in the image domain. In the new image, the centers oL the hsxagon3 have high magnitude. The actual mag-nitude of th~ ~pots at th2 hexagon centers i~ depsndent on how 4 many edgoo w~xo 1A it~ neighborhood. More ~dg~s create greater f en~rgy at allowed r~usncle3 and he~ce high magnitude spots.
Fewer edg~s give ri~ to lower magnituda ~pots. Tha magnitude of ths ~pot~ i9 a good mea~ure of the confidonce leval in the clock re~toration at any given poiAt.

(~) Malor A~ termina~lon Tha hexagonal lmage ha3 now be~n r~cr~at~d but its ori~ntation nead~ to be dotarmined.
Ih~ hexagonal honeycomb pattern of tha invention has threo "axas" 3paced 60 degreas apart. Th~ direction of ~hese axeo 18 e~tablished by t~e brighte~ poin~i in tha transform domain ater spa~iaI bandpass fil~erin~. It i3 now possible to a~certain which o thesa three axss i~ the ma~or axi3. This step ,,., ...: . :

~ 32~3 is optional. If this step i5 not performed the label would haveto be decoded ~hree times, using each of the three axeq, with only one axis yielding a meaningful message. The maior axi~ is arbitrarily chosen a3 the~ axis which runs parallel to two sides of the label as described hetreinabove and depicted in Fig. 2.
If the boundaries of the square labol are determined based on the knowledge of th~ major axi-s, then most of the energy in the restored hexagonal outline pattern will be inside this 3quar~'q boundaries.
To determine the major axl~ ~ach of the three axas is assumetd to be the ma~or axis. The conseguent 3quara label out-li~e i9 determlned for each trial axi3, and tho total clock resto-ration pattarn en~rgy interior to that square is detarmined fro~
th~ dlgital ene,rgy dat~ output from the inverse trantform subrou-tine. The correct trial i8 the onet with the most enorgy. The angle of thls ma~or axls is then stored for the lnitialization step and other soarching ope~rations. At thls ~unctura, it ls not yet known whether the rccorded angle i~ in the correct direction or 180 degr~e away from the correct direction. ~he ~ource code 20 li~ting~ ln the appo~d~d Mlcroftche Appendix pertaining to the -d~torminatlon of tha ma~or axi~ may be ~ound at pago 49, lines 48 to 54; pa~o 50, line~ 1 to 53; and paga 51, lin~ l to 5. It will be appreciated that all three label areas do not need to be de~t~rmined ln toto, since the anargy in tha areas common to all three squarea does not neted to be detorminad.
~) Saaxchin~
A ~to~od progr~m antltl~d "S~RC~.C" 300 d~picted on ~ig. 10 combin-a th~ Tran ~rmod and raganeratod h~xagon center in~ormation wlth th~ ~tort3d intenslty l~v~ls of the origi~al image 90 as to determine the gray level value o~ each hexagon.
Th~ search is performad in such a way aq to minimize the chances of "getting lost" while searchlng. The end re ult i to obtain a matrix o~ th~ gray level value for aach hexago~ of th~ data array.
rhe sourcs codes listings for "SEARCH.C" may b~ found within the , 132~9~3 Microfiche Appendix at page 52 through 60. Foux important in-formatlon arrayY are constructed during the first part of the SEARC~.C program. The array CVAL (clock value) stores a measure of the quality of the recovered clock 3ignal for each hexagon, whila tho array GVA~ sto~ tho g~oy level valu~ ~0-63) at the center of each hexagon. The remaining arrays IVAL and JVAL store the row and column locatlon~ of the center or aach hexagon.

(i) Initialization S~
From the major axi angle determined in step (e) and the known spacing of the h~xagon~ (5 pixels) in the example, the expocted horizontal and vertical displacement~ from tha center of on~ hexagon to the center3 of the ~urrounding ~ix hexagon~ are computed.
Following these cGmputationq, the SEARCH.C pro~ram operate~ on the clock recovery slgnal, retrieved from memory and the re~caled labal imago, also rotrievad from momory. The funda-mental purpQse o tha initialization subroutine found in the Micro ich~ Appendlx at page 52, line~ 13 to 54; paga 53, lines 1 to 48; pags 56, line~ 47 to 57; and page 57 line~ 1 to 35 is to msr~ a~d condenso tho information from tho3~ two 30urce~ and to generato a da~a matrix provlding tho grey 3c~1e value for each hexagon.
Thè initialization step of the s~arch i~ bounded by a ~quaro around the label'~ center o about 1/3 of an lnch on a 3ide. Wlthin this area, a ~ood starting point ls ~he polnt of highest magn~tud~ in t~e racov~red clock 3ignal array is fo~nd.
Than, th~ location of this 3tarting polnt relative to the center o~ tha labol iY determinad. Thi~ Ytarting point i~ a point where the clook ~lgnal i~ ~trong and di~tinct, and also a point rela-tiv~ly near tha canter of the labal. A strong, distinct signal is de~ired to en~ure that saarching begin~ wlth a valid hexagon center, and it is deqired that the point be near the center of the label ~o that its a~olute locat~on can be dater~lned without seriou~ influence from warping or tilting. The measure of the ::
"

' ' : ~' ' 132~

quality of a point in the clocX recovery pattern is the point's magnitude minuq the magnitude of it3 eight ~urrounding points.
The rectangular coordinates of the starting point are converted to polar form, the polar coordinate~ are adjusted relative to the previously determlned maior axis angle, and this reRult is con-verted back to rectangular form. These coordinates are scaled according to the expected row spacing (4.5 pixels) and column spacing (5 pixels) to arrive at the insertion po~ition on the hexagon matrix. The clock guality, grey level and locations corre~ponding to the starting hexagon axa then inserted in the respective arrays CVAL, GVAL, IV~L and JVA~.

~ ain Sear~h Loop The main qearch loop proceeds to locate the centers of the remaining hexagons. Tha Loop tarminates when the expected number of hexagons has been located. The order o the ~earch for haxagon centars i~ extromaly important. ~h~ incroa~ad reliability of th~ d~codlng proce~3 in tha faca o~ labol degradations comes from the part1cular search technigu~ employ~d, as de#crlbed ~elow.
Each itaratlon of the search loop begins by recalling th~ locatiOn of the hlghest magnitude clo~k recovary spot whose neighbor~ have not been 3earched for their strongest values.
Erom thl~ known point, th~ earch will b~ extanded one hexagon i~
each o~ six direction The effect i8 to build up the qearch patter~ along a path from better to worso recovored clock qual-ity. Thus, 1 there is a woak aroa of r~covered clock, ~.~. at ths label c~nter or an oblit~rated area, the search algorithm s~irts around it rather ~ha~ going through it. By out1anking these weak ar~as and 3aving th~m for last, the probability of getting lost o~ the grid i~ greatly reduced. Sinca getting lost is ju3t a3 bad a~ rea~ins a gray level lncorractly, this charac-teristic o the ~earch algorithm i~ extremely powarful.
A subroutine ound in the Mlcrofiche Appendix at page 53, line~ S0 to 54; page 54, linas 1 to 53; and page 55, lines 1 35 to 55, i3 responsibla for searching the neighbor~ oi~ the best ' ' ' ~ , - . :

. . , " 1 3 ~ S~
quality clock value found in the maln loop. The subroutine loops six times, once for each hexagonal neighbor of the hexagon then under con3ideration. First, the po~ition of a nelghbor is com-puted. If this neighbor i~ outsida the label boundary, the loop lteration terminateq. If not, the neighbor i~ checkad to see if it has already been searched ~rom another direction. The loop iteration will terminate if the neighbor has been ~earched, since the algorithm maXes earlier ~earches more rellable than later one~. rf th2 naighbor get~ b~yond this te~t, the expected po-~it$on o~ thH noighbor' 9 csnter ln the clocX recovery pattern iscomputed. At this point, a gradient search for the hicJh~st magni-tud6 clocX ~iynal i~ perform~d. The eight pix019 surrou~d~ng the recovered position are ~earchad to see i~ a higher clocX value is found. If it is, then tho best neighboring point has its eight neighbors checked to soe lf an even bettar value is found. This gradient search provido~ a d~grea of adaptatlon which is impera-tivs lf warped and tllted labels ara to be read. The ~ubroutine then ~oe9 on to the next nolghbor or retu~ns when all neighbors hava been checked.
As montioned abova und~r ~tep (d), aq a result of the da~a tran~ormation proc~s~es, the racon~tructed grid now carries in~ormation regardi~g the geometric centers o~ the polygonal ~ncodin~ celI~. Thi3 grid has m~r~ ~n~rgy in area~ where more contra~ting intorfaca~ originalIy oxlsted. Th~ centers will 119 on tho predet0rmln~d two-dim3n~10nal array having a predeterminad numbor o~ e~ually- or unegually-spaced axes, as the case may be.
Th3 informatlon concern~ng tha spa~ial rela~ion~hip of the axes of the pradete~mined two-dimensional array may de~irably be used i~ the ma~or axi~ orlentat~on step.
It will b~ appr~ciata~, however, that tha al~orithm could bo approprlatoly modifIed to hava the decoding proce~s datermlne tho actual geometry o the two-dimen~ional array and from that doterminatlon proceed to determino the ~lltration - schem0, the ao-called ma~or axi~i of the label fi.e the axis of the two-dimen~ional array that ii3 parallel to two sides of a , ' . ': ' ' ~;,"' ` ~ ' ':
, ' , , : .

9 ~ 3 square optically readable label as discussed herein) and provide the n~ce~sary coordinates for the searching subroutine.
Whethar the geometry of the label is determined by such an optional step a9 described abova or s~imply entered into the decoding proce~s through appropriate moclifications to ~he two-dimensional clock recovery process, t~a variety of label con-figuration~ disclosed and discus~ed herein can be easily accommo-dat3d by one o ordinary skill in the art. It will be appreclated that the numbcr of axos upon which the c:enter~ of the individual, ad~acent polygonal encoding cells are arrayed and their respective angular orlentation, can be substituted in the major axi deter-mination tep for the threa axes of tho hexagonal array of the preerred embodiment. Theraore the major axis of the predeter-~ined two-dimensional array can be determin~d wlthout performing the trial and srror ~naly~i~ do~cribed above in qtep (e).
A~ for the hexagonal array of tho preferred embodiment, tha lnformatlon from th~ ma~or axls determination step and the known spacing of polyqons may be used ~o compute the expected horizontal and v~rtical displacem~nt~ from tho canter of one polygo~ to the co~ter~ of surrounding polygons. Following these computation~ and after making the nocessary ad~ustments to the ~carch ~ubroutin~, tha ooarch, includ'lng ths initialization step ~nd maln s~arch loop 8top ca~ proceed for th~ particular label co~iguration that ~ being employod. It will ba apprcciated that such ~inor ad~ustments to the search routin2 SEARCH.C 300 in tho app~nded ~ource coda listings are within the abllitie~ of a p~rson of ordi~ary skill tn the art.
: After the subroutine complete~, tho current center lo-catio~ i5 mark~d ~o that it i~ not sear~h~d again. The e~fect is to d~l~te thi~ po3~ion as a candidate for having its neighbor~
~ea~chod. For each loop iteration, from O to 6 new candidates are added and on~ candidate is deleted. An ~fficient implemen-tation could u~ a data structure which keep~ candidateY in magni tude order as inYert and delet~ oparation~ are performed. One such ~tructura is called a priority ~ueue (Refer~nce: Tbe De_~g3 ~: :

1 3 ~ 3 ", .
an~ AnalY~ of Com~u~er Alaorithms, Aho, Hopcroft and Ullman, (Addison We~ley, 1974)). It i3 Xnown that a linear search algo-rithm reguires order n2 operation3 wherea~ an efficiPntly imple-mented priority queue using a balanced trea or heap structure requires order n log n operationq. An order n search algorithm based on bucket sorting could also be used, if recovered clock value-e2 are scaled and redu ed to a small range of intsg~rs.

~ isto~ram Generation and Thre~holdi~
After the main 3~arch loop terminates, the locations of the center~ of all hexagon~ have been dotermlned and the gray values of the centerY of all hexagon , which hav~ been ltored, are completely illed in. The next tep iq to threshold the dig-itized grey level v~lua3 in tho range 0 - 63 to the discrete level~ of, for exampleh black, grey, and whlte (for a black, white and g~9y label). Thi3 i~ done by building a hie~togram of the label image intensity value~ from the hexagon centars. Slicing l~vel3 can be determined by looXing for dip~ in the histogram.
Th~ ~pecific subroutine utilized to co~truct the hie~togram and d~termin~ thc ~licing levels may be found in the app~nded source code li~ting~ in the Microfiche Appendix at pagc S5, lines 16 to 52 and pag~ 56, lines 1 to 15.
' ~
(h) Co~rso G~ld ~orroc~iQn_and Final Orient~tlon A ter thre~holding to di3crete leval~, two di.~tortions may ~till be prosent. Fir~t, the array may be of center. Thi~
can happen i~ tho inl~ial 3earch step doe~ not correctly deter-mi~e th~ locatlon of the b~ t qu31ity clock ~i~nal reLative to tha lab~l c~nt~r. The 3~0nd po3sibllity i3 that th~ encire lab~l has effectively boen read upslde down qince tha major axis angle has an ambiguity o~ 180 degrees.
A 3tored subroutine found at page 5~, lines 1 to 54 and page 59, line~ 1 to 24 within the ~l~ro~iche ~ppendix performs the function of de~ermi~ing whether t~e label i8 o~f center. If thc label ha3 been positioned correctly, the coordinates o the 132~3~3 center row ~hould pass through the center o the label. To deter-mine if a vertical positioning error has been ~ade, rows above the hypothesized center row are checked to see which would orm a lina passing clo~est to the Label center. If a row above or below 5 is closer than the hypothesized center row, then the appropriate shift up or down is made. If the left justi~ication of short row9 ha~ been performed incorrectly, thi 3 i9 ad~u~ted by shifting ~hort rows one po~ition to the right.
aorizontal positioning errore and up~ide down reading are checked using inormation embedd~d in the label known a~
coarYe grid in~ormation. The in~ormation i8 di~tributed in 3 cell x 3 ccll clu~ter~ of hexagons a3 described hereinabove.
S~nce the label may he, for example, on a 33 row by 30 column grid, tho~a clu~te~s form a 11 by 10 grid. The bottom center hexagon o~ each complete 3 cell x 3 cell cluster has a spsciaL
property whlch i~ provided during encodlng. ~here 1~ a guaran-teed transitlon on either side of this hexagon, as previou~ly described in connection with Fig. 4. For exampl~, if ths bottom center hexagon ie black, the bottom left and bottom right hexa-gon~ mu~t bo olthor gray or white. A ~torad subroutinc found atpa~o 59, lin~e 27 to 52 and pag~ 60, linoa 1 to 33 o~ the Micro-fich~ App~ndix take~ advantage of this transitlon property to ra~ove the 1nal two possible distortion~. Fir~t ~n array is created where ~ach element of th~ array indicate3 whether a tran~ition took plac~ betweon two horizontally ad~ac~nt hexagons.
Then the array i~ checked for each of 9 hypothetical slides o~
the coar30 grid arrangod a~ a 3 x 3 pattorn around the expected sl~d~ of 0. Ono oY the~3 slide~ will ~how a b~tt~r match between actual and expecte~ tran3it~0ns, and thi~ ~;id~ posltion is re-tain~d. Next, th~ ~amu hypo~hesl~ ia checked under the as~ump-tion that thR label was ~ead upside down. Thi~ will happen if th~ ma~or axis angle actually pointed right ~o laf~ in relation to how the label was printed rather than let to right.
I~ the label was Qimply invertad, i~., lnterchanged higher row with lower rows and higher columns with lower col-umna, th2n th9 results of sliding~ shoul~ be inverted a~ well.

, , 13299~3 Howev~r, one important trans ormation must be performed to cor-rectly inv~rt the label. During reading the qhort (length 29) rowq are left ju3tified; thu-, when the label i~ i~verted these label~ mu3t be right justifled. The adju~tmont i~ made, and it 5 i9 thl~ proc~dur~ which will make the r~sults of the slide hypo-the~es other than a simple inversion. In fact, the best result from the slide te~ts will be better than any previous test i~ the label wa3 actually read up~ide down.
~aving determined whether or not the label was read up3id~ down, and whether there wa~ any lide in the absolute po~itioning, the label matrix can now be decoded. Wlth correct dot~rmination o th~ imag~ and lido, the image proc3~ing func-tlon~ ar~ complete and the data decodlng proce~o~ ar~ qtarted.

4. Dacodlna A stored program "RD.LABEL.C" 182 on Flg. 9 found withln withln the Mlcro~lcho Appendix at paga 61, llne~ 1 to 52, and pag~ 62, llne~ 1 to 2~ read3 the flle ganerated by the ~earch program and genqratos a bit stream fila with, in the pro~or~od ambodlment, 1292 bi~. It u~e~ a ~tored subroutine Cell Dec.C 183 on Elg. 9 and contained ln th~ Microfiche Appendix at pagos 63 through 66 to ma3~ out unusabla hexa~on~, and to apply d~codlng whlch i3 th~ in~er~a o~ the codlng pso~ram.
The ir~t ~tep in tho decoding proc~ to generate a 25 bit ~tream ~rom the hoxagon information, using a hoxagon to-bit mappl~g procas~ which i~ the rovor3e of th~ blt-to-hexaqon map-ping proc~s~ us~d ln tho ~ncoding opsratlon.
Tho blt ~information) str~am is then bifurcated by the program into a hlgh priority r.a~age bit ~tr~am and a low 1 30 prlority m~ssago blt~3tream or a~ many bit ~treams a3 are used in '~ encodln~ the la~ol.
~ It i~ then necessary to apply error correc~ion to each i bit stream using the error coding technigue~ which we~e used in the Label encoding proce~s. For exampl~, i ~eed-Solomon coding iq used, error correction on the bit stream generated by the .. , . ... ~., . 1' ,, ~32~53 search program generates an output which is in the same format as previously de~cribed for the encoding input file. Error correction may be per ormed in the following 3aquence (Roference: ~k~
and_eractic.e o~ ~rror ControL Code~, de3cribed above.) S
1. Compute syndrome~
. Computo Error Loca~or Po:Lynomial u~ing Berlekamp-Mass~y Algoritlln 3. Compute error locationq,u~ing Chien search 4. Computa ~rror mag~itudeA u~ing Eor~oy'3 Also~ithm The la~t ~top i9 ax~cuted only-lf a corr~ctable numb~r of error~
has b~e~ detected from 3teps 2 and 3. rho number o~ error d~-- tected are al30 computed. If an uncorrectable nu~ber of errors i~ detected or an error i9 located in the lmplied padding (des-crlbed abovo), a ~lag i~ ~et. The Sp2Ci~iC error coding proce-dure utllized ln tha illustrativ~ exampl,e may be found in the Micro icho Appondix at page 67 through 75, and i9 deqlgnated as ERRDEC.C la4 on Fig. 9.
.

5.. Q~E~
By trackin~ th~ package (by identi~ying its location on th~ convcyor) th~ high priori~y mes~ g~, indicating the zip code ~.
o~ tho packago d~tlnatto~ can ba used to activate ~uitable rou~-ing arm~ or conveyor~ to route the package to th~ propcr truck, alrplane or pacXag~ carrier to tako the packago to i~9 dos~ination.
Al~hough th~ invon~ion may b~ a~ us~d in a conveyor/
dlvertor ~y~tam, it will bo apparont that lt can b~ u3ad in a wid~ vari~ty o~ in ormation gathering, package handling and pro-duction operatior~ in whlch it i~ de~ired to road a label on a packaga, letter, part, machin~ or the llXe and cau~e a 3y9tem ~0 perform a package handling or manufactuxing operation, or example, on ~he o~ject bearing the labal. Tha invantion allows the~ oparations to occur with high speed, hiyh a~curacy, dealinq with a ~ubs~antial amount of iabel information and even protact-ing much o~ that information rom being los~ due to labal tearsand tha lika.

, .

~ ~329~3 With reference to Fig. 9, to alternatively display the decoded me~sage on a computer tarminal, the prog~am TEXTOUT.C 185 may be amployed. Program TEXTOUT.C may be found within the Micro- :
fiche Appendix at page~ 76 through 78.

:
: 25 30.

. , ~ ~ ; :

, : :

~. : ; . ,

Claims (171)

1. An optically readable label for storing encoded information comprising a multiplicity of information-encoded hexagons contiguously arranged in a honeycomb pattern, each hexagon having one of at least two different optical properties.

2. An article as recited in Claim 1, wherein said optical properties, are the colors black, white and gray.

3. An article as recited in Claim 1, wherein more important information is encoded in hexagons proximate the center of said article.

4. An article as recited in Claim 1, wherein the information encoded in said hexagons includes at least a first and second message area and said first message area is located farther from the periphery of said article than said second message area.

5. An article as recited in Claim 1, wherein said information-encoded hexagons are encoded with message infor-mation and error detection information, thereby allowing errors in the message information retrieved from said article to be detected.

6. An article as recited in Claim 5, wherein said error detection information may be utilized to correct errors in the message information retrieved from said article.

7. An article as recited in Claim 1, further com-prising a plurality of Concentric Rings occupying an area on said article separate from the area occupied by said information-encoded hexagons, each Concentric Ring having one of at least two different optical properties in alternating sequence.

8. An article as recited in Claim 7, wherein said Concentric Rings are centrally located on said article.

9. An article as recited in Claim 8, wherein said contiguous information-encoded hexagons are arranged in up to about fifty rows and up to about fifty columns within an area of up to about one square inch.

10. An article as recited in Claim 8, wherein said contiguous information-encoded hexagons are arranged in up to about thirty-three rows and up to about thirty columns within an area of up to about one square inch and wherein said Concentric Rings occupy less than about ten percent of the area of said article.

11. An article as recited in Claim 7, wherein the information encoded in said hexagons includes at least a first and second message area and said first message area is located farther from the periphery of said article than said second message area.

12. An article as recited in Claim 7, wherein the Concentric Rings occupy less than about twenty-five percent of the area of said article.

13. An article as recited in Claim 7, wherein more important information is encoded in hexagons proximate to the center of said article.

14. An article as recited in Claim 7, wherein said optical properties of said hexagons are the colors black, white and gray.

15. An article as recited in Claim 14, wherein the optical properties of said Concentric Rings are the same as two of the two or more optical properties of said hexagons.

16. An article as recited in Claim 15, wherein the optical properties of said Concentric Rings are alternately black and white.

17. An optically readable label for storing encoded information comprising a multiplicity of contiguously arranged, information-encoded polygons other than squares or rectangles, each polygon having one of at least two different optical prop-erties.

18. An article as recited in Claim 17 further comprising a plurality of Concentric Rings on said article, each Concentric Ring alter-nately having one of at least two different optical properties.

19. An article as recited in Claim 18, wherein said Concentric Rings are centrally located on said article.

20. A process for encoding information in an optically-readable label comprising a multiplicity of information-encoded hexagons contiguously arranged in a honeycomb pattern, each hexa-gon having one of at least two optical properties, comprising the steps of:
(a) assigning one of at least two optical prop-erties to each hexagon to create a plurality of contiguous hexa-gons having different optical properties;
(b) encoding the information by ordering the hexagons in a predetermined sequence; and (c) printing each hexagon in its assigned opti-cal property.

21. A process as recited in Claim 20, further com-prising the steps of:
(d) assigning a plurality of dots in a dot matrix to define the optical property of each hexagon; and (e) printing said plurality of dots.

22. A process as recited in Claim 20, wherein step (b) includes the step of mapping groups of two or more contig-uous hexagons in predetermined geographical areas on said article.

23. A process as recited in Claim 21, wherein step (b) includes the step of mapping groups of two or more contiguous hexagons in predetermined geographical areas on said article.

24. A process as recited in Claims 22 or 23, further comprising the steps of dividing the information being encoded into at least two categories of higher and lower priorities, and encoding said higher and lower priority information in separate, predetermined geographical areas.

25. A process as recited in Claim 20, further com-prising the step of encoding a plurality of selected hexagons with error detection information and interposing said error detection encoded hexagons among said hexagons.

26. A process as recited in Claim 24, further com-prising the step of encoding a plurality of selected hexagons with error detection information and interposing said error detection encoded hexagons among said hexagons.

27. A process as recited in Claim 26, wherein sep-arate encoded error detection information is separately applied to said higher and lower priority information.

28. A process as recited in Claim 25, wherein said encoded error detection information may be utilized to correct errors in the information retrieved from said article.

29. A process as recited in Claim 27, wherein said error detection information may be utilized to correct errors in the information retrieved from said article.

30. A process as recited in Claims 20 or 21, wherein said encoding step is structured to optimize the number of con-tiguous hexagons having different optical properties.

31. A process of storing and retrieving data, com-prising the steps of:

(a) printing on a label a multiplicity of information-encoded hexagons contiguously arranged in a honeycomb pattern, each hexagon having one of at least two different optical properties;
(b) illuminating said label;
(c) optically sensing light reflected from said hexagons with an electro-optical sensor;
(d) generating analog electrical signals corresponding to the intensities of light reflected from said optical properties as sensed by individual pixels of said sensor;
(e) converting said analog electrical signals into sequenced digital signals;
(f) storing said digital signals in a storage medium connected to a computer to form a replica of said digital signals in said storage medium;
(g) decoding said replica of said digital signals to retrieve the characteristics of the intensities, locations and orientations of the individual optical properties of said hexagons; and (h) generating a digital bit stream output from the computer representing the decoded information represented by the hexagons.

32. A process as recited in Claim 31, wherein said optical properties are the colors black and white.

33. A process as recited in Claim 31, wherein said optical properties are the colors black, white and gray.

34. A process as recited in Claim 31, further comprising the step of normalizing the stored digital signals to predetermined digital signal levels corresponding to said optical properties.

35. A process as recited in Claim 31, wherein said hexagons are encoded by performing the steps of:
(i) assigning one of at least two optical properties to each hexagon to create a plurality of contiguous hexagons having different optical properties;
(ii) encoding the information by ordering the hexagons in a predetermined sequence; and (iii) printing each hexagon in its assigned optical property.

36. A process as recited in Claim 31, wherein said hexagons are encoded by performing the steps of:
(i) assigning one of at least two optical properties to each hexagon to create a plurality of contiguous hexagons having different optical properties;
(ii) encoding the information by ordering the hexagons in a predetermined sequence and mapping groups of two or more contiguous hexagons in predetermined geographical areas on said article; and (iii) printing each hexagon in its assigned optical property.

37. A process of storing and retrieving data, comprising the steps of:
(a) printing on a label a multiplicity of information-encoded hexagons contiguously arranged in a honeycomb pattern, and a plurality of centrally-located Concentric Rings, each hexagon having one of at least two different optical properties and said Concentric Rings having alternating optical properties corresponding to at least two of the optical properties of said hexagons;
(b) illuminating said label;

(c) optically sensing light reflected from said hexagons and said Concentric Rings with an electro-optical sensor;
(d) generating analog electrical signals corresponding to the intensities of light reflected from said hexagons and said Concentric Rings as sensed by individual pixels of said sensor;
(e) filtering said analog electrical signals through an analog bandpass filter to determine the presence of said Concentric Rings, thereby detecting the presence of said hexagons within the field of view of said sensor;
(f) converting said analog electrical signals into a sequenced digital bit stream;
(g) storing said digital signals in a storage medium to form a replica of said digital signals in said storage medium;
(h) decoding said replica of said digital signals to retrieve the characteristics of the intensities, locations and orientations of the individual optical properties of said hexagons; and (i) generating a digital bit stream output from said computer representing the decoded hexagons.

38. A process as recited in Claim 37, wherein the optical properties of said hexagons and said Concentric Rings are the colors black and white.

39. A process as recited in Claim 37, wherein the optical properties of said hexagons are the colors black, white and gray and the optical properties of said Concentric Rings are the same as two of the optical properties of said hexagons.

40. A process as recited in Claim 37, further comprising the step of normalizing said stored data to predetermined digital signals corresponding to said optical properties of said hexagons.

41. A process as recited in Claim 37, wherein said hexagons are encoded by performing the steps of:
(i) assigning one of at least two optical properties to each hexagon to create a plurality of contiguous hexagons having different optical properties;
(ii) encoding the information by ordering the hexagons in a predetermined sequence; and (iii) printing each hexagon in its assigned optical property.

42. A process as recited in Claim 37, wherein said hexagons are encoded by performing the steps of:
(i) assigning one of at least two optical properties to each hexagon to create a plurality of contiguous hexagons having different optical properties;
(ii) encoding the information by ordering the hexagons in a predetermined sequence and mapping groups of two or more contiguous hexagons in predetermined geographical areas on said article; and (iii) printing each hexagon in its assigned optical property.

43. A process of storing and retrieving data, comprising the steps of:
(a) printing on a substrate a multiplicity of information-encoded hexagons contiguously arranged in a honeycomb pattern, and a plurality of centrally-located Concentric Rings, each hexagon having one of at least two different optical properties, and said Concentric Rings having alternating optical properties corresponding to at least two of the optical properties of said hexagons;
(b) illuminating said substrate;

(c) optically sensing light reflected from said hexagons and said Concentric Rings with an electro-optical sensor;
(d) transmitting digital electrical signals corresponding to the intensity of light reflected from said hexagons and said Concentric Rings as recorded by individual pixels of said sensor;
(e) filtering said digital electrical signals through a digital bandpass filter to determine the presence of said Concentric Rings, thereby detecting the presence of said hexagons within the field of view of said sensor;
(f) storing said digital electrical signals in a storage medium connected to a computer to form a replica of said digital electrical signals in said storage medium;
(g) decoding said replica of said digital elec-trical signal to retrieve the characteristics of the intensities, locations and orientations of the individual optical properties of said hexagons; and (h) transmitting a digital bit stream output from said computer representing the decoded hexagons.

44. A process as recited in Claim 43, wherein salt digital bandpass filter is a two-dimensional digital bandpass filter.

45. A process for decoding a stream of digital sig-nals representing an electro-optically sensed image correspond-ing to a multiplicity of contiguously-arranged polygons encoded in a predetermined pattern, each polygon having one of at least two optical properties, comprising the steps of:
(a) performing a two-dimensional clock recovery on said image to determine the coordinates and intensities of said optical properties;
(b) searching said intensities of the optical properties of step (a) to identify the optical properties of said contiguously-arranged polygons; and (c) decoding said polygons by performing the inverse of the encoding process for said polygons.

46. A process as recited in Claim 45, wherein said contiguously arranged polygons are hexagons arranged in a honeycomb pattern.

47. A process as recited in Claim 45, wherein step (b) comprises:
(i) performing an initilization step which searches the two-dimensional clock recovered coordinates and intensities of said optical properties determined in step (a) within a prede-termined area of said multiplicity of polygons to identify the position of greatest intensity; and (ii) performing a search continuation loop step which searches the two-dimensional clock recovered coordinates and intensities from the position of greatest intensity in step (i) and looping to each adjacent position of next greatest in-tensity, wherein each identified position corresponds to the center of a polygon.

48. A process as recited in Claim 46, wherein step (b) comprises:
(i) performing an initialization step which searches the two-dimensional clock recovered coordinates and intensities of the optical properties determined in step (a) within a prede-termined area of said image, to identify the position of greatest intensity; and (ii) performing a search continuation loop step which searches the two-dimensional clock recovered coordinates and intensities of said optical properties over the entire image starting from the position of greatest intensity in step (i) and looping to each adjacent position of next greatest in-tensity, wherein each identified position corresponds to the center of a hexagon.

49. A process as recited in Claim 45, wherein step (a) comprises the steps of:
(i) performing a non-linear mapping operation on said digital signals to identify transitions between adjacent polygons having different optical properties;
(ii) performing a Fourier transformation on the non-linear mapped digital signals to obtain two-dimensional, non-linear coordinate corresponding to the direction, spacing and intensity of optical property transitions of said polygons;
(iii) filtering said two-dimensional non-linear coordinates to eliminate incorrect direction and spacing of optical property transitions of said polygons; and (iv) performing an inverse Fourier transforma-tion on said filtered two-dimensional non-linear coordinates to restore digital signals corresponding to a replicated image of said polygons recorded by said electro-optical sensor.

50. A process as recited in Claim 49, wherein said polygons are hexagons contiguously-arranged in a honeycomb pattern.

51. A process as recited in Claim 49, wherein step (i) comprises creating a two-dimensional map of the transitions between adjacent polygons having different optical properties by computing the standard deviation of the optical properties of said image recorded by each pixel and pixels proximate each pixel of said electro-optical sensor, wherein larger standard deviation value correspond to transition areas at the inter-faces of said polygons.

52. A process as recited in Claim 47, further com-prising the step of thresholding said transformed digital sig-nals corresponding to the center of each polygon located in step (ii) to determine the respective optical properties of said polygons.

53. A process as recited in Claim 52, wherein the step of determining the thresholds of said transformed digital signals is performed by constructing histograms representing the respective optical properties of said polygons.

54. A process as recited in Claim 45, further com-prising the step, prior to step (a), of normalizing the sensed image to predetermined levels for each respective optical prop-erty of the image.

55. A process as recited in Claim 45, further comprising the step, prior to step (a), of rescaling said image to create an image with equal horizontal and vertical magnification.

56. A process as recited in Claim 50, further com-prising the step of determining the major axis of said hexagons by first determining all of the axes of said hexagons and then determining which of these axes has a predetermined relationship to a boundary of the image.

57. A process as recited in Claim 49, further com-prising the step, before performing the Fourier transformation step, of windowing the non-linear mapped digital signals to reduce the intensities of optical properties sensed by said electro-optical sensor which are not associated with said polygons.

58. A process as recited in Claim 49, wherein said image sensed by said electro-optical sensor includes an acqui-sition target comprising a plurality of Concentric Rings of different, alternating optical properties and wherein the first step of the process is locating said acquisition target by fil-tering said digital signals and correlating said digital sig-nals to a signal of predetermined frequency.

59. A combination optical mark sensing and decoding system, comprising:
(a) an optically readable label for storing en-coded data comprising a multiplicity of information-encoded hexagons contiguously arranged in a honeycomb pattern, each hexagon having one of at least two different optical properties;
(b) means for illuminating a predetermined area;
(c) means for optically imaging said predeter-mined illuminated area through which said label is arranged to pass and generating analog electrical signal corresponding to the intensities of light reflected from said hexagons and striking each pixel of said imaging means;
(d) means for converting said analog electrical signals into a sequenced digital bit stream corresponding to the intensities of light recorded by said pixels of said imaging means;
(e) means for storing said digital bit stream for subsequent decoding of said label; and (f) means for decoding said digital bit stream, said decoding means producing an electrical output representa-tive of the encoded information.

60. An apparatus as recited in Claim 59, wherein said optically readable label further comprises a plurality of Concentric Rings, said Concentric Rings having alternating op-tical properties corresponding to at least two of the optical properties of said hexagons.

61. An apparatus as recited in Claim 60, wherein said Concentric Rings are centrally located on said label.

62. An apparatus as recited in Claim 61, wherein each hexagon is black, white or gray and said Concentric Rings are alternating black and white.

63. An apparatus as recited in Claim 60, further comprising means for filtering said analog electrical signals to determine the presence of said Concentric Rings, thereby de-tecting the presence of said label within said predetermined illuminated area.

64. An apparatus as recited in Claims 59 or 60, wherein said optical imaging means comprises a charged coupled device.

65. An optical mark sensing and decoding system for an optically readable label for storing encoded data comprising a multiplicity of information-encoded hexagons contiguously ar-ranged in a honeycomb pattern, each hexagon having one of at least two different optical properties, comprising:
(a) means for illuminating a predetermined area;
(b) means for optically imaging said predeter-mined illuminated area through which said label is arranged to pass and generating analog electrical signals corresponding to the intensities of light reflected from said hexagons and striking each pixel of said imaging means;

(c) means for converting said analog electrical signals into a sequenced digital bit stream corresponding to the intensities of light recorded by said pixels of said imaging means;
(d) means for storing said digital bit stream for subsequent decoding of said label; and (e) means for decoding said digital bit stream, said decoding means producing an electrical output representa-tive of the encoded information.

66. An optical mark sensing and decoding system for an optically readable label for storing encoded data comprising a multiplicity of information-encoded hexagons contiguously arranged in a honeycomb pattern and a plurality of centrally-located Concentric Rings, each hexagon having one of at least two different optical properties and said Concentric Rings hav-ing alternating optical properties corresponding to at least two of the optical properties of said hexagons; comprising;
(a) means for illuminating a predetermined area;
(b) means for optically imaging said predeter-mined illuminated area through which said label is arranged to pass and generating analog electrical signals corresponding to the intensities of light reflected from said hexagons and striking each pixel of said imaging means;
(c) means for converting said analog electrical signals into a sequenced digital bit stream corresponding to the intensities of light recorded by said pixels of said imaging means;
(d) means for storing said digital bit stream for subsequent decoding of said label; and (e) means for decoding said digital bit stream, said decoding means producing an electrical output representa-tive of the encoded information.

67. An apparatus as recited in Claim 66, further comprising means for filtering said analog electrical signals to determining the presence of said Concentric Rings, thereby de-tecting the presence of said label within said predetermined illuminated area.

68. An apparatus for decoding a stream of digital signals representing an electro-optically sensed image corre-sponding to a multiplicity of contiguously-arranged polygons encoded in a predetermined pattern, each polygon having one of at least two optical properties, comprising:
(a) means for performing a two-dimensional clock recovery on said image to determine the coordinates and intensities of said optical properties;
(b) means for searching said intensities of the optical properties of step (a) to identify the optical proper-ties of said polygons; and (c) means for decoding said polygons by per-forming the inverse of the encoding process for said polygons.

69. An apparatus for decoding a stream of digital signals representing an electro-optically sensed image of a multiplicity of contiguously-arranged polygons encoded in a predetermined pattern and each polygon having one of at least two optical properties, comprising:
(a) means for performing a non-linear mapping operation on said digital signals to identify transitions be-tween adjacent polygons having different optical properties;
(b) means for performing a Fourier transforma-tion on the non-linear mapped digital signals to obtain a two-dimensional map corresponding to the direction, spacing and intensity of optical property transitions of said polygons;

(c) means for filtering said two-dimensional map to eliminate incorrect direction and spacing of optical property transitions of said polygons;
(d) means for performing an inverse Fourier transformation on said filtered two-dimensional map to restore digital signals corresponding to a replicated image of said polygons;
(e) means for searching the transformed digital signals to determine the optical property of the center of each polygon and its location within said multiplicity of polygons;
and (f) means for decoding said polygons by perform-ing the inverse of the encoding process for said polygons.

70. Apparatus as recited in Claim 69, wherein means (e) comprises:
(i) initialization means to search said trans-formed digital signals within a predetermined area of said image to identify the position of greatest intensity; and (ii) search continuation loop means to search said transformed digital signals over the entire image starting from the position of greatest intensity in means (i) and looping to each adjacent position of next greatest intensity, wherein each identified position corresponds to the center of a polygon.

71. Apparatus as recited in Claims 69 or 70, wherein said polygons are hexagons contiguously-arranged in a honeycomb pattern.

72. Apparatus as recited in Claim 69, wherein said non-linear mapping means comprises means for creating a two-dimensional map of the transitions between adjacent polygons having different optical properties by computing the standard deviation of the optical properties of said image recorded by each pixel and pixels proximate each pixel of said electro0 optical sensor, wherein larger standard deviation values cor-respond to transition areas at the interfaces of said polygons.

73. Apparatus as recited in Claim 69, further com-prising means for thresholding said transformed digital sig-nals corresponding to the center of each polygon located by means (e) to determine the respective optical properties of said polygons.

74. Apparatus as recited in Claim 73, wherein the thresholding means comprising means for constructing histograms representing the respective optical properties of said polygons.

75. Apparatus as recited in Claim 69, further com-prising means for normalizing the sensed image to predetermined optimums for each respective optical property of the image prior to performing said non-linear mapping operation.

76. Apparatus as recited in Claim 75, further comprising means for rescaling the normalized image, to create an image with equal horizontal and vertical magnifica-tion prior to performing said non-linear mapping operation.

77. Apparatus as recited in Claim 71, further com-prising means for determining the major axis of said hexagons by first determining all of the axes of said hexagons and then determining which of these axes has a predetermined relationship to a boundary of the image.

78. Apparatus as recited in Claim 69, further com-prising means for windowing the non-linear mapped digital signals to reduce the intensities of optical properties sensed by said electro-optical sensor which are not associated with said polygons before performing said Fourier transformation on the non-linear mapped digital signals.

79. Apparatus as recited in Claim 69, wherein said image sensed by said electro-optical sensor includes an acqui-sition target comprising a plurality of Concentric Rings of dif-ferent, alternating optical properties and means for locating said acquisition target by filtering said digital signals and correlating said digital signals to a signal of predetermined frequency.

80. An optically readable label for storing encoded information comprising a multiplicity of information-encoded polygons having at least five sides, said polygons arranged with the geometric centers of adjacent polygons lying at the vertices of a predetermined two-dimensional array, said polygons having one of at least two different optical properties.

81. An article as recited in Claim 80, wherein said array is a hexagonal array.

82. An article as recited in claim 81, wherein said hexagonal array has three axes spaced 60 degrees apart.

83. An article as recited in claim 80, wherein said polygons are substantially in the shape of regular hexagons.

84. An article as recited in claim 80, wherein said optical properties are the colors black, white and gray.

85. An article as recited in claim 80, wherein said polygons are irregular polygons.

86. An article as recited in claims 80 or 81, further comprising a plurality of Concentric Rings occupying an area on said article separate from the area occupied by said information-encoded polygons, each Concentric Ring having one of at least two different optical properties in alternating sequence.

87. An article as recited in claim 86, wherein said Concentric Rings are centrally located on said article.

88. An optically readable label for storing encoded information comprising a multiplicity of information-encoded triangles, said triangles arranged with the geometric centers of adjacent triangles lying at the vertices of a predetermined two-dimensional array, and said triangles having one of at least two different optical properties.

89. An article as recited in claim 88, further compris-ing a plurality of Concentric Rings occupying an area on said article separate from the area occupied by said information-encoded triangles, each Concentric Ring having one of at least two different optical properties in alternating sequence.

90. An article as recited in claim 89, wherein said Concentric Rings are centrally located.

91. An optically readable label for storing encoded information comprising a multiplicity of information-encoded pollygons, said polygons arranged with the geometric centers of adjacent polygons lying at the vertices of a two-dimensional hexagonal array, and said polygons having one of at least two different optical properties.

92. An article as recited in claim 91, wherein said polygons are substantially in the shape of regular hexagons.

93. An optically readable label for storing encoded information comprising a multiplicity of information-encoded polygons, said polygons arranged with the geometric centers of adjacent polygons lying at the vertices of a predetermined two-dimensional array, and said polygons having one of at least two different optical properties and said array having at least three equally-spaced axes.

94. An optically readable label for storing encoded information comprising a multiplicity of information-encoded polygons, said polygons partially contiguously arranged with the geometric centers of adjacent polygons lying at the vertices of a predetermined two-dimensional array, and said polygons having one of at least two different optical properties.

95. An optically readable label for storing encoded information comprising a multiplicity of information-encoded poly-gons, said polygons noncontiguously arranged with the geometric centers of adjacent polygons lying at the vertices of a predeter-mined two-dimensional array, and said polygons having one of at least two different optical properties.

96. An article as recited in claims 93, 94 or 95, wherein said polygons are regular polygons.

97. An article as recited in claims 93, 94 or 95, wherein said polygons are irregular polygons.

98. An article as recited in claims 93, 94 or 95, further comprising a plurality of Concentric Rings occupying an area on said article separate from the area occupied by said information-encoded polygons, each Concentric Ring having one of at least two different optical properties in alternating sequence.

99. An article as recited in claim 98, wherein said Concentric Rings are centrally located on said article.

100. An article as recited in claims 94 or 95, wherein said array is a hexagonal array.

101. An article as recited in claim 100, wherein said hexagonal array has three axes spaced 60 degrees apart.

102. A process for decoding a stream of digital signals representing an electro-optically sensed label image correspond-ing to a multiplicity of noncontiguously-arranged polygons en-coded in accordance with an encoding process, said polygons de-fining a multiplicity of interstitial spaces among said polygons, said polygons arranged with the geometric centers of adjacent polygons lying at the vertices of a predetermined two-dimensional array, and said polygons and said interstitial spaces having one of at least two different optical properties, comprising the steps of:
(a) performing a two-dimensional clock recovery on said sensed label image to obtain a recovered clock signal;
(b) utilizing said recovered clock signal of step (a) to locate the geometric centers of said polygons to identify the optical properties of said polygons; and (c) decoding said polygons by performing the inverse of said encoding process.

103. A process for decoding a stream of digital signals representing an electro-optically sensed label image corresponding to a multiplicity of partially contiguously-arranged polygons encoded in accordance with an encoding process, said polygons defining a multiplicity of interstitial spaces among said polygons, said polygons arranged with the geometric centers of adjacent polygons lying at the vertices of a predetermined two-dimensional array, and said polygons and said interstitial spaces having one of at least two different optical properties, comprising the steps of:
(a) performing a two-dimensional clock recovery on said sensed label image to obtain a recovered clock signal;
(b) utilizing said recovered clock signal of step (a) to locate the geometric centers of said polygons to identify the optical properties of said polygons; and (c) decoding said polygons by performing the in-verse of said encoding process.

104. A process as recited in claims 102 or 103, wherein said two-dimensional array is a hexagonal array.

105. A process as recited in claims 102 or 103, wherein said polygons are regular polygons.

106. A process as recited in claims 102 or 103, wherein said polygons are irregular polygons.

107. A process as recited in claims 102 or 103, wherein said polygons are substantially in the shape of regular hexagons.

108. A process as recited in claim 102, wherein step (a) comprises the steps of:
(i) performing a nonlinear mapping operation on said digital signals to identify transitions between adjacent polygons and between polygons and interstitial spaces having different optical properties;
(ii) performing a Fourier transformation on the nonlinear mapped digital signals to obtain a two-dimensional rep-resentation corresponding to the direction, spacing and intensity of optical property transitions of said polygons;
(iii) filtering said transformed nonlinear mapped digital signals to eliminate incorrect direction and spacing of optical property transitions of said polygons; and (iv) performing an inverse Fourier transformation on said filtered transformed nonlinear mapped digital signals to obtain said recovered clock signal.

109. A process as recited in claim 102, further compris-ing the step, prior to step (a), of normalizing the sensed label image to predetermined levels for each respective optical property of the image.

110. A process as recited in claim 102, further compris-ing the step, prior to step (a) of rescaling the image to create an image with equal horizontal and vertical magnification.

111. A process as recited in claim 108, wherein step (i) comprises creating a two-dimensional map of the transitions between adjacent polygons and between polygons and said inter-stitial spaces having different optical properties by computing the standard deviation of the optical properties of said image recorded by each pixel and pixels proximate each pixel of said electro-optical sensor, wherein larger standard deviation values correspond to transition areas at the interfaces of said polygons.

112. A process as recited in claim 111, further compris-ing the step of thresholding said sensed label image at the center of each polygon located in step (b) to determine the respective optical properties of said polygons.

113. A process as recited in claim 112, wherein the step of determining the thresholds of said sensed label image is per-formed by constructing histograms representing the respective optical properties of said polygons.

114. A process as recited in claims 102, 108 or 113, wherein step (b) comprises:
(i) performing an initialization step which searches the two-dimensional recovered clock signal obtained in step (a) within a predetermined area of said signal, to identify the position of greatest intensity; and (ii) performing a search continuation loop step which searches the two-dimensional recovered clock signal over the entire recovered clock signal starting from the position of greatest intensity in step (i) and looping to each adjacent position of next greatest intensity, wherein each identified position corresponds to the center of a polygon.

115. A process as recited in claim 114, wherein said image sensed by said electro-optical sensor includes an acqui-sition target comprising a plurality of Concentric Rings of different, alternating optical properties and wherein the first step of the process is locating said acquisition target by fil-tering said digital signals and correlating said digital signals to a signal of predetermined frequency.

116. A process as recited in claim 103, wherein step (a) comprises the steps of:
(i) performing a nonlinear mapping operation on said digital signals to identify transitions between adjacent polygons and between polygons and said interstitial spaces having different optical properties;
(ii) performing a Fourier transformation on the nonlinear mapped digital signals to obtain a two-dimensional rep-resentation corresponding to the direction, spacing and intensity of optical property transitions of said polygons;
(iii) filtering said transformed nonlinear mapped digital signals to eliminate incorrect direction and spacing of optical property transitions of said polygons; and (iv) performing an inverse Fourier transformation on said filtered transformed nonlinear mapped digital signals to obtain said recovered clock signal.

117. A process as recited in claim 103, further compris-inq the step, prior to step (a), of normalizing the sensed label image to predetermined levels for each respective optical prop-erty of the image.

118. A process as recited in claim 103, further compris-ing the step, prior to step (a) of rescaling the image to create an image with equal horizontal and vertical magnification.

119. A process as recited in claim 116, wherein step (i) comprises creating a two-dimensional map of the transitions be-tween adjacent polygons and between polygons and said interstitial spaces having different optical properties by computing the stan-dard deviation of the optical properties of said image recorded by each pixel and pixels proximate each pixel of said electro-optical sensor, wherein larger standard deviation values corres-pond to transition areas at the interfaces of said polygons.

120. A process as recited in claim 119, further compris-ing the step of thresholding said sensed label image at the center of each polygon located in step (b) to determine the respective optical properties of said polygons.

121. A process as recited in claim 120, wherein the step of determining the thresholds of said sensed label image is per-formed by constructing histograms representing the respective optical properties of said polygons.

122. A process as recited in claims 103, 116 or 121, wherein step (b) comprises:
(i) performing an initialization step which searches the two-dimensional recovered clock signal obtained in step (a) within a predetermined area of said signal, to identify the position of greatest intensity; and (ii) performing a search continuation loop step which searches the two-dimensional recovered clock signal over the entire recovered clock signal starting from the position of greatest intensity in step (i) and looping to each adjacent posi-tion of next greatest intensity, wherein each identified position corresponds to the center of a polygon.

123. A process as recited in claim 122, wherein said image sensed by said electro-optical sensor includes an acquisi-tion target comprising a plurality of Concentric Rings of dif-ferent, alternating optical properties and wherein the first step of the process is locating said acquisition target by filtering said digital signals and correlating said digital signals to a signal of predetermined frequency.

124. A combination optical mark sensing and decoding system, comprising:
(a) an optically readable label for storing en-coded information comprising a multiplicity of information-encoded polygons having at least five sides, said polygons ar-ranged with the geometric centers of adjacent polygons lying at the vertices of a predetermined two-dimensional array, and said polygons having one of at least two different optical properties;
(b) means for illuminating a predetermined area;
(c) means for optically imaging said predeter-mined illuminated area through which said label is arranged to pass and generating analog electrical signals corresponding to the intensities of light reflected from said polygons and strik-ing each pixel of said imaging means;
(d) means for converting said analog electrical signals into a sequenced digital bit stream corresponding to the intensities of light recorded by said pixels of said imaging means;
(e) means for storing said digital bit stream for subsequent decoding of said label; and (f) means for decoding said digital bit stream, said decoding means producing an electrical output representative of the encoded information.

125. An apparatus as recited in claim 124, wherein said optically readable label further comprises a plurality of Concen-tric Rings, said Concentric Rings having alternating optical prop-erties corresponding to at least two of the optical properties of said polygons.

126. A combination optical mark sensing and decoding system, comprising:
(a) an optically readable label for storing en-coded information comprising a multiplicity of information-encoded polygons, said polygons arranged with the geometric centers of adjacent polygons lying at the vertices of a two-dimensional hexagonal array, and said polygons having one of at least two different optical properties;
(b) means for illuminating a predetermined area;
(c) means for optically imaging said predeter-mined illuminated area through which said label is arranged to pass and generating analog electrical signals corresponding to the intensities of light reflected from said polygons and strik-ing each pixel of said imaging means;
(d) means for converting said analog electrical signals into a sequenced digital bit stream corresponding to the intensities of light recorded by said pixels of said imaging means;
(e) means for storing said digital bit stream for subsequent decoding of said label; and (f) means for decoding said digital bit stream, said decoding means producing an electrical output representative of the encoded information.

127. An apparatus as recited in claim 126, wherein said optically readable label further comprises a plurality of Concen-tric Rings, said Concentric Rings having alternating optical prop-erties corresponding to at least two of the optical properties of said polygons.

128. An apparatus as recited in claim 127, wherein said polygons are substantially in the shape of a regular hexagon.

129. A combination optical mark sensing and decoding system, comprising:
(a) an optically readable label for storing en-coded information comprising a multiplicity of information-encoded triangles, said triangles arranged with the geometric centers of adjacent triangles lying at the vertices of a predetermined two-dimensional array, and said triangles having one of at least two different optical properties;
(b) means for illuminating a predetermined area;
(c) means for optically imaging said predeter-mined illuminated area through which said label is arranged to pass and generating analog electrical signals corresponding to the intensities of light reflected from said triangles and strik-ing each pixel of said imaging means;
(d) means for converting said analog electrical signals into a sequenced digital bit stream corresponding to the intensities of light recorded by said pixels of said imaging means;
(e) means for storing said digital bit stream for subsequent decoding of said label; and (f) means for decoding said digital bit stream, said decoding means producing an electrical output representative of the encoded information.

130. An apparatus as recited in claim 129, wherein said optically readable label further comprises a plurality of Concen-tric Rings, said Concentric Rings having alternating optical prop-erties corresponding to at least two of the optical properties of said polygons.

131. A combination optical mark sensing and decoding system, comprising:
(a) an optically readable label for storing en-coded information comprising a multiplicity of information-encoded polygons, said polygons noncontiguously-arranged with the geometric centers of adjacent polygons lying at the vertices of a predeter-mined two-dimensional array, said polygons having one of at least two different optical properties;
(b) means for illuminating a predetermined area;
(c) means for optically imaging said predeter-mined illuminated area through which said label is arranged to pass and generating analog electrical signals corresponding to the intensities of light reflected from said polygons and striking each pixel of said imaging means;
(d) means for converting said analog electrical signals into a sequenced digital bit stream corresponding to the intensities of light recorded by said pixels of said imaging means;

(e) means for storing said digital bit stream for subsequent decoding of said label; and (f) means for decoding said digital bit stream, said decoding means producing an electrical output representative of the encoded information.

132. An apparatus as recited in claim 131, wherein said optically readable label further comprises a plurality of Concen-tric Rings, said Concentric Rings having alternating optical properties corresponding to at least two of the optical properties of said polygons.

133. A combination optical mark sensing and decoding system, comprising:
(a) an optically readable label for storing en-coded information comprising a multiplicity of information-encoded polygons, said polygons partially contiguously-arranged with the geometric canters of adjacent polygons lying at the vertices of a predetermined two-dimensional array, said polygons having one of at least two different optical properties;
(b) means for illuminating a predetermined area;
(c) means for optically imaging said predeter-mined illuminated area through which said label is arranged to pass and generating analog electrical signals corresponding to the intensities of light reflected from said polygons and strik-ing each pixel of said imaging means;
(d) means for converting said analog electrical signals into a sequenced digital bit stream corresponding to the intensities of light recorded by said pixels of said imaging means;
(e) means for storing said digital bit stream for subsequent decoding of said label; and (f) means for decoding said digital bit stream, said decoding means producing an electrical output representative of the encoded information.

134. An apparatus as recited in claim 133, wherein said optically readable label further comprises a plurality of Concen-tric Rings, said Concentric Rings having alternating optical prop-erties corresponding to at least two of the optical properties of said polygons.

135. An apparatus for decoding a stream of digital sig-nals representing an electro-optically sensed label image of a multiplicity of noncontiguosly-arranged polygons encoded in ac-cordance with an encoding process, said polygons defining a mul-tiplicity of interstitial spaces among said polygons, said poly-gons arranged with the geometric centers of adjacent polygons lying at the vertices of a predetermined two-dimensional array, and said polygons and said interstitial spaces having one of at least two different optical properties, comprising:
(a) means for performing a two-dimensional clock re-covery on said sensed label image to obtain a recovered clock signal;
(b) means for utilizing said recovered clock signal of step (a), to locate the geometric centers of said polygons and identify the optical properties of said polygons; and (c) means for decoding said polygons by performing the inverse of said encoding process.

136. An apparatus for decoding a stream of digital sig-nals representing an electro-optically sensed label image of a multiplicity of noncontiguously-arranged polygons encoded in ac-cordance with an encoding process, said polygons defining a multi-plicity of interstitial spaces among said polygons, said polygons arranged with the geometric centers of adjacent polygons lying at the vertices of a predetermined two-dimensional array, and said polygons and said interstitial spaces having one of at least two different optical properties, comprising:
(a) means for performing a nonlinear mapping operation on said digital signals to identify transitions between adjacent polygons having different optical properties;
(b) means for performing a Fourier transformation on the nonlinear mapped digital signals to obtain a two-dimensional representation corresponding to the direction, spacing and in-tensity of optical property transitions of said polygons;
(c) means for filtering said transformed non-linear mapped digital signals to eliminate incorrect direction and spacing of optical property transitions of said polygons;
(d) means for performing an inverse Fourier transformation of said filtered transformed nonlinear mapped digital signals to obtain said recovered clock signal;
(e) means for utilizing said recovered clock signal to locate the geometric center of said polygons and to identify the optical properties of said polygons; and (f) means for decoding said polygons by perform-ing the inverse of said encoding process for said polygons.

137. An apparatus as recited in claim 136, wherein said nonlinear mapping means comprises means for creating a two-dimensional map of the transitions between adjacent polygons having different optical properties by computing the standard deviation of the optical properties of said image recorded by each pixel and pixels proximate each pixel of said electro-optical sensor, wherein larger standard deviation values corres-pond to transition area at the interfaces of polygons.

138. An apparatus as recited in claim 136, further com-prising means for normalizing the sensed label image to predeter-mined optimums for each respective optical property of the image prior to performing the nonlinear mapping operation.

139. An apparatus as recited in claim 136, further com-prising means for rescaling the sensed label image to create an image with equal horizontal and vertical magnification prior to performing the nonlinear mapping operation.

140. An apparatus as recited in claim 137, further com-prising means for thresholding said sensed label image at the center of each polygon located by means (e) to determine the re-spective optical properties of said polygons.

141. An apparatus as recited in claim 140, wherein the thresholding means further comprises means for constructing histograms representing the respective optical properties of said polygons.

142. An apparatus as recited in claims 135, 136 or 141, wherein said searching means comprises:
(i) initialization means to search said two-dimensional recovered clock signal within a predetermined area of said signal, to identify the position of greatest intensity; and (ii) a search continuation loop means to search said two-dimensional recovered clock signal over the entire re-covered clock signal starting from the position of greatest in-tensity obtained by means (i) and looping to each adjacent posi-tion of next greatest intensity, wherein each identified position corresponds to the center of a polygon.

143. An apparatus as recited in claim 142, wherein said image sensed by said electro-optical sensor includes an acquisi-tion target comprising a plurality of Concentric Rings of dif-ferent, alternating optical properties and means for locating said acquisition target by filtering said digital signals and correlating said digital signals to a signal of predetermined frequency.

144. An apparatus for decoding a stream of digital signals representing an electro-optically sensed label image of a multiplicity of partially contiguously-arranged polygons encoded in accordance with an encoding process, said polygons defining a multiplicity of interstitial spaces among said polygons, said polygons arranged with the geometric centers of adjacent polygons lying at the vertices of a predetermined two-dimensional array, and said polygons and said interstitial spaces having one of at least two different optical properties, comprising:
(a) means for performing a two-dimensional clock recovery on said sensed label image to obtain a recovered clock signal;

(b) mean for utilizing said recovered clock signal to locate the geometric centers of said polygons and identify the optical properties of said polygons; and (c) means for decoding said polygons by perform-ing the inverse of said encoding process.

145. An apparatus for decoding a stream of digital signals representing an electro-optically sensed label image of a multiplicity of partially contiguously-arranged polygons defining a multiplicity of interstitial spaces among said polygons, said polygons arranged with the geometric centers of adjacent polygons lying at the vertices of a predetermined two-dimensional array, said polygons and said interstitial spaces having one of at least two different optical properties, comprising:
(a) means for performing a nonlinear mapping operation on said digital signals to identify transitions between adjacent polygons having different optical properties;
(b) means for performing a Fourier transformation on the nonlinear mapped digital signals to obtain a two-dimensional representation corresponding to the direction, spac-ing and intensity of optical property transitions of said poly-gons;
(c) means for filtering said transformed non-linear mapped digital signals to eliminate incorrect direction and spacing of optical property transitions of said polygons;
(d) means for performing an inverse Fourier transformation of said filtered transformed non-linear mapped digital signals to obtain said recovered clock signal;
(e) means for utilizing said recovered clock signal to locate the geometric centers of said polygons and identify the optical properties of said polygons; and (f) means for decoding said polygons by performing the inverse of said encoding process for said polygons.

146. An apparatus as recited in claim 145, wherein said nonlinear mapping means comprises means for creating a two-dimensional map of the transitions between adjacent polygons hav-ing different optical properties by computing the standard devia-tion of the optical properties of said image recorded by each pixel and pixels proximate each pixel of said electro-optical sensor, wherein larger standard deviation values correspond to transition areas at the interfaces of polygons.

147. An apparatus as recited in claim 145, further com-prising means for normalizing the sensed label image to predeter-mined optimums for each respective optical property of the image prior to said nonlinear mapping operation.

148. An apparatus as recited in claim 145, further com-prising means for rescaling the sensed label image to create an image with equal horizontal and vertical magnification prior to said nonlinear mapping operation.

149. An apparatus as recited in claim 146, further com-prising means for thresholding said sensed label image at the center of each polygon located by means (e) to determine the re-spective optical properties of said polygons.

150. An apparatus as recited in claim 149, wherein the thresholding means further comprises means for constructing his-tograms representing the respective optical properties of said polygons.

151. An apparatus as recited in claims 144, 145 or 150, wherein said searching means comprises:
(i) initialization means to search said two-dimensional recovered clock signal within a predetermined area of said signal to identify the position of greatest intensity; and (ii) a search continuation loop means to search said two-dimensional recovered clock signal over the entire re-covered clock signal from the position of greatest intensity ob-tained by means (i) and looping to each adjacent position of next greatest intensity, wherein each identified position corresponds to the center of a polygon.

152. An apparatus as recited in claim 151, wherein said image sensed by said electro-optical sensor includes an acquisi-tion target comprising a plurality of Concentric Rings of dif-ferent, alternating optical properties and means for locating said acquisition target by filtering said digital signals and correlating said digital signals to a signal of predetermined frequency.

153. A process for encoding information in an optically readable label comprising a multiplicity of partially contiguously arranged polygons defining a multiplicity of in-terstitial spaces among said polygons, said polygons arranged with the geometric centers of adjacent polygons lying at the vertices of a predetermined two-dimensional array, and said polygons and said interstitial spaces having one of at least two different optical properties, comprising the steps of:
(a) assigning one of at least two optical prop-erties to each polygon to create a plurality of partially contiguously-arranged polygons having different optical proper-ties;
(b) encoding the information by ordering the polygons in a predetermined sequence; and (c) printing each polygon in its assigned optical property.

154. A process as recited in claim 153, further compris-ing the steps of:
(d) assigning a plurality of dots in a dot matrix to define the optical property of each polygon; and (e) printing said plurality of dots.

155. A process for encoding information in an optically readable label comprising a multiplicity of contiguously-arranged polygons, said polygons arranged with the geometric centers of adjacent polygons lying at the vertices of a predetermined two-dimensional array, said polygons having one of at least two dif-ferent optical properties, comprising the steps of:
(a) assigning one of at least two optical prop-erties to each polygon to create a plurality of contiguously-arranged polygons having different optical properties;
(b) encoding the information by ordering the poly-gons in a predetermined sequence; and (c) printing each polygon in its assigned optical property.

156. A process as recited in claim 155, further compris-ing the steps of:
(d) assigning a plurality of dots in a dot matrix to define the optical property of each polygon; and (e) printing said plurality of dots.

157. A process for encoding information in an optically readable label comprising a multiplicity of noncontiguously-arranged polygons defining a multiplicity of interstitial spaces among said polygons, said polygons arranged with the geometric centers of adjacent polygons lying at the vertices of a predeter-mined two-dimensional array, and said polygons and said intersti-tial spaces having one of at least two different optical proper-ties, comprising the steps of:
(a) assigning one of at least two optical prop-erties to each polygon to create a plurality of noncontiguously-arranged polygons having different optical properties;
(b) encoding the information by ordering the poly-gons in a predetermined sequence; and (c) printing each polygon in its assigned optical property.

158. A process as recited in claim 157, further compris-ing the steps of:
(d) assigning a plurality of dots in a dot matrix to define the optical property of each polygon; and (e) printing said plurality of dots.

159. A process as recited in claims 153, 155 or 157, wherein step (b) includes the step of mapping groups of two or more polygons in predetermined geographical areas on said article.

160. A process as recited in claim 159, further compris-ing the steps of dividing the information being encoded into at least two categories of higher and lower priorities, and encoding said higher and lower priority information in separate, predeter-mined geographical areas.

161. A process as recited in claim 160, further compris-ing the step of separately applying error detection information to said higher and lower priority information.

162. A process as recited in claims 153, 155 or 157, fur-ther comprising the step of encoding a plurality of selected polygons with error detection information and interposing said error-detection-encoded polygons among said polygons.

163. A process as recited in claim 161, further compris-ing the step of utilizing said error detection information to correct errors in the information retrieved from said article.

164. A process as recited in claim 162, wherein said error detection information may be utilized to correct errors in the information retrieved from said article.

165. A process as recited in claims 153, 155 or 157, fur-ther comprising the step of structuring said encoding step to optimize the number of polygons having different optical prop-erties.

166. A process of storing and retrieving data, compris-ing the steps of:
(a) printing on a label a multiplicity of par-tially contiguously-arranged polygons encoded in accordance with an encoding process said polygons defining a multiplicity of interstitial spaces among said polygons, said polygons arranged with the geometric centers of adjacent polygons lying at the vertices of a predetermined two-dimensional array, and said polygons and said interstitial spaces having one of at least two different optical properties;
(b) illuminating said label;
(c) optically sensing light reflected from said polygons with an electro-optical sensor;
(d) generating analog electrical signals cor-responding to the intensities of light reflected from said op-tical properties as sensed by individual pixels of said sensor;
(e) converting said analog electrical signals into sequenced digital signals;

(f) storing said digital signals in a storage medium connected to a computer to form a replica of said digital signals in said storage medium;
(g) decoding said replica of said digital signals to retrieve the characteristics of the intensities, locations and orientations of the individual optical properties of said poly-gons; and (h) generating a digital bit stream output from the computer representing the decoded information represented by the polygons.

167. A process as recited in claim 166, wherein said label further comprises a plurality of centrally-located Concen-tric Rings, said Concentric Rings having alternating optical properties corresponding to at least two of the optical proper-ties of said polygons.

168. A process of storing and retrieving data, compris-ing the steps of:
(a) printing on a label a multiplicity of noncontiguously-arranged polygons encoded in accordance with an encoding process, said polygons defining a multiplicity of in-terstitial spaces among said polygons, said polygons arranged with the geometric centers of adjacent polygons lying at the vertices of a predetermined two-dimensional array, and said polygons and said interstitial spaces having one of at least two different optical properties;
(b) illuminating said label;
(c) optically sensing light reflected from said polygons with an electro-optical sensor;
(d) generating analog electrical signals cor-responding to the intensities of light reflected from said op-tical properties as sensed by individual pixels of said sensor;
(e) converting said analog electrical signals into sequenced digital signals;

(f) storing said digital signals in a storage medium connected to a computer to form a replica of said digital signals in said storage medium;
(g) decoding said replica of said digital signals to retrieve the characteristics of the intensities, locations and orientations of the individual optical properties of said poly-gons; and (h) generating a digital bit stream output from the computer representing the decoded information represented by the polygons.

169. A process as recited in claim 168, wherein said label further comprises a plurality of centrally-located Con-centric Rings, said Concentric Rings having alternating optical proper-ties corresponding to at least two of the optical properties of said polygons.

170. A process of storing and retrieving data, compris-ing the steps of:
(a) printing on a label a multiplicity of contiguously-arranged polygons encoded in accordance with an en-coding process, said polygons arranged with the geometric centers of adjacent polygons lying at the vertices of a predetermined two-dimensional array, and said polygons and said interstitial spaces having one of at least two different optical properties, (b) illuminating said label;
(c) optically sensing light reflected from said polygons with an electro-optical sensor;
(d) generating analog electrical signals cor-responding to the intensities of light reflected from said op-tical properties as sensed by individual pixels of said sensor;
(e) converting said analog electrical signals into sequenced digital signals;
(f) storing said digital signals in a storage medium connected to a computer to form a replica of said digital signals in said storage medium;

(g) decoding said replica of said digital signals to retrieve the characteristics of the intensities, locations and orientations of the individual optical properties of said poly-gons; and (h) generating a digital bit stream output from the computer representing the decoded information represented by the polygons.

171. A process as recited in claim 170, wherein said label further comprises a plurality of centrally-located Concen-tric Rings, said Concentric Rings having alternating optical properties corresponding to at least two of the optical proper-ties of said polygons.